Model profile · OpenAI

GPT-5.4 mini

Closed weightspremium · registry tag 2026 mini

Visible tradeoffs

Reads as visible tradeoffs across the resolved source data.

Visible coverage: 63%
Verified coverage: 55.5%
Spread: 90%
Last verified: Jun 20, 2026

42%bench fit

textcodevisiondocumentsearch19 aliases46 official source links

Open compare

Data version

Current snapshot.

Data version Jun 20, 2026Model list checked9 providers · 1081 tracked modelsPage refreshed Jul 5, 2026

The registry snapshot and page stamp are shown so a stale deploy is visible at a glance.

Source-linked scores by benchmark

Each row keeps the benchmark source, source type, raw metric, and percentile inside its fair comparison set.

Visible tradeoffsThis model currently reads as visible tradeoffs across the resolved source data.

Chat / text38 benchmarks48.5%

Intelligence Index

AA · Chat / text · Combined

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #171 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 14
Percentile: 57%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `intelligenceIndex`.

57% percentile inside its fair comparison set

14Raw benchmark value

AA-Omniscience accuracy

AA · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #136 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 18.2%
Percentile: 54.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceAccuracy`.

54.7% percentile inside its fair comparison set

18.2%Raw benchmark value

AA-Omniscience non-hallucination

AA · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #276 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 4.7%
Percentile: 7.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceNonHallucination`.

7.7% percentile inside its fair comparison set

4.7%Raw benchmark value

IFBench

AA · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #182 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 38.8%
Percentile: 42.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `ifbench`.

42.5% percentile inside its fair comparison set

38.8%Raw benchmark value

Blended price

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #199 · Source label: GPT-5.4 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $1.7 /1M tokens
Percentile: 28.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mBlended0To3To1`.

28.3% percentile inside its fair comparison set

$1.7 /1M tokensRaw benchmark value

Input price

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #189 · Source label: GPT-5.4 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $0.8 /1M input tokens
Percentile: 31.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mInputTokens`.

31.9% percentile inside its fair comparison set

$0.8 /1M input tokensRaw benchmark value

Output price

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #206 · Source label: GPT-5.4 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $4.5 /1M output tokens
Percentile: 25.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mOutputTokens`.

25.7% percentile inside its fair comparison set

$4.5 /1M output tokensRaw benchmark value

Output Speed

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #105 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 94.6 tokens/s
Percentile: 50.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianOutputTokensPerSecond`.

50.5% percentile inside its fair comparison set

94.6 tokens/sRaw benchmark value

Time to first token

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #207 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 104.11s
Percentile: 1.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstTokenSeconds`.

1.9% percentile inside its fair comparison set

104.11sRaw benchmark value

Time to first answer token

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #206 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 104.11s
Percentile: 2.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstAnswerTokenSeconds`.

2.4% percentile inside its fair comparison set

104.11sRaw benchmark value

Openness Index

AA · Chat / text · Combined

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #183 · Source label: GPT-5 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 6
Percentile: 7.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `opennessBreakdown.opennessIndex`.

7.5% percentile inside its fair comparison set

6Raw benchmark value

Text Arena

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #39 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,448
Percentile: 88.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #50. Votes: 39525. Organization: openai. License: Proprietary.

88.3% percentile inside its fair comparison set

1,448Raw benchmark valueCI 1,444 - 1,452

Text Arena · Creative Writing

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #61 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,399
Percentile: 81.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing. Source rank: #75. Votes: 6391. Organization: openai. License: Proprietary.

81.4% percentile inside its fair comparison set

1,399Raw benchmark valueCI 1,390 - 1,407

Text Arena · English

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #42 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,452
Percentile: 87.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #54. Votes: 18792. Organization: openai. License: Proprietary.

87.4% percentile inside its fair comparison set

1,452Raw benchmark valueCI 1,446 - 1,457

Text Arena · Exclude Ties

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #37 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,450
Percentile: 88.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: exclude_ties. Source rank: #48. Votes: 30409. Organization: openai. License: Proprietary.

88.9% percentile inside its fair comparison set

1,450Raw benchmark valueCI 1,445 - 1,456

Text Arena · Hard Prompts

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #37 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,471
Percentile: 88.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts. Source rank: #49. Votes: 25673. Organization: openai. License: Proprietary.

88.9% percentile inside its fair comparison set

1,471Raw benchmark valueCI 1,466 - 1,476

Text Arena · Hard Prompts English

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #43 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 87%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts_english. Source rank: #55. Votes: 12804. Organization: openai. License: Proprietary.

87% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,462 - 1,475

Text Arena · Instruction Following

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #44 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 86.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: instruction_following. Source rank: #55. Votes: 13187. Organization: openai. License: Proprietary.

86.8% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,426 - 1,439

Text Arena · Longer Query

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #48 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,446
Percentile: 84.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: longer_query. Source rank: #61. Votes: 16352. Organization: openai. License: Proprietary.

84.5% percentile inside its fair comparison set

1,446Raw benchmark valueCI 1,440 - 1,453

Text Arena · Multi Turn

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #25 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,472
Percentile: 92.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: multi_turn. Source rank: #34. Votes: 7385. Organization: openai. License: Proprietary.

92.6% percentile inside its fair comparison set

1,472Raw benchmark valueCI 1,464 - 1,480

Text Arena · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #81 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,412
Percentile: 75.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #97. Votes: 39525. Organization: openai. License: Proprietary.

75.4% percentile inside its fair comparison set

1,412Raw benchmark valueCI 1,408 - 1,416

Text Arena · Creative Writing · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #85 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,367
Percentile: 74%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing. Source rank: #105. Votes: 6391. Organization: openai. License: Proprietary.

74% percentile inside its fair comparison set

1,367Raw benchmark valueCI 1,359 - 1,375

Text Arena · English · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #93 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,413
Percentile: 71.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #109. Votes: 18792. Organization: openai. License: Proprietary.

71.7% percentile inside its fair comparison set

1,413Raw benchmark valueCI 1,407 - 1,418

Text Arena · Exclude Ties · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #81 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 75.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: exclude_ties. Source rank: #96. Votes: 30409. Organization: openai. License: Proprietary.

75.4% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,393 - 1,404

Text Arena · Hard Prompts · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #69 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,426
Percentile: 79.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts. Source rank: #84. Votes: 25673. Organization: openai. License: Proprietary.

79.1% percentile inside its fair comparison set

1,426Raw benchmark valueCI 1,421 - 1,431

Text Arena · Hard Prompts English · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #82 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,422
Percentile: 75%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts_english. Source rank: #98. Votes: 12804. Organization: openai. License: Proprietary.

75% percentile inside its fair comparison set

1,422Raw benchmark valueCI 1,415 - 1,428

Text Arena · Instruction Following · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #66 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,403
Percentile: 80%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: instruction_following. Source rank: #80. Votes: 13187. Organization: openai. License: Proprietary.

80% percentile inside its fair comparison set

1,403Raw benchmark valueCI 1,397 - 1,409

Text Arena · Longer Query · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #80 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,407
Percentile: 74%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: longer_query. Source rank: #98. Votes: 16352. Organization: openai. License: Proprietary.

74% percentile inside its fair comparison set

1,407Raw benchmark valueCI 1,401 - 1,413

Text Arena · Multi Turn · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #48 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,434
Percentile: 85.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: multi_turn. Source rank: #60. Votes: 7385. Organization: openai. License: Proprietary.

85.4% percentile inside its fair comparison set

1,434Raw benchmark valueCI 1,427 - 1,442

Instruction following

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #102 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 18.9%
Percentile: 6.5%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: IF. Tasks scored: 4.

6.5% percentile inside its fair comparison set

18.9%Raw benchmark value

Language

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #105 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 41.8%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Language. Tasks scored: 3.

3.7% percentile inside its fair comparison set

41.8%Raw benchmark value

Paraphrase

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 14.6%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: paraphrase. Category: IF.

4.6% percentile inside its fair comparison set

14.6%Raw benchmark value

Simplify

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #92 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 22.9%
Percentile: 15.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: simplify. Category: IF.

15.7% percentile inside its fair comparison set

22.9%Raw benchmark value

Story generation

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #102 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 16.8%
Percentile: 6.5%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: story_generation. Category: IF.

6.5% percentile inside its fair comparison set

16.8%Raw benchmark value

Summarize

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 20.2%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: summarize. Category: IF.

4.6% percentile inside its fair comparison set

20.2%Raw benchmark value

Connections

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #100 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 48.5%
Percentile: 8.3%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: connections. Category: Language.

8.3% percentile inside its fair comparison set

48.5%Raw benchmark value

Plot unscrambling

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 25.9%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: plot_unscrambling. Category: Language.

4.6% percentile inside its fair comparison set

25.9%Raw benchmark value

Typos

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 46%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: typos. Category: Language.

3.7% percentile inside its fair comparison set

46%Raw benchmark value

Coding25 benchmarks49.9%

Terminal-Bench Hard

AA · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #128 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 14.4%
Percentile: 57.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `terminalbenchHard`.

57.9% percentile inside its fair comparison set

14.4%Raw benchmark value

SciCode

AA · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #99 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 36.9%
Percentile: 73.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `scicode`.

73.4% percentile inside its fair comparison set

36.9%Raw benchmark value

Coding Index

AA · Coding · Combined

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #17 · Source label: GPT-5.4 mini (xhigh)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 56
Percentile: 78.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `codingIndex`.

78.7% percentile inside its fair comparison set

56Raw benchmark value

Agentic Index

AA · Coding · Combined

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #13 · Source label: GPT-5.4 mini (xhigh)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 30
Percentile: 73.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `agenticIndex`.

73.9% percentile inside its fair comparison set

30Raw benchmark value

Code Arena

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 57.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #40. Votes: 7713. Organization: openai. License: Proprietary.

57.5% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,391 - 1,406

WebDev Arena

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 57.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: webdev. Source rank: #40. Votes: 7713. Organization: openai. License: Proprietary.

57.5% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,391 - 1,406

Code Arena · Webdev Html

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #30 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,416
Percentile: 60.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: webdev-html. Source rank: #38. Votes: 960. Organization: openai. License: Proprietary.

60.3% percentile inside its fair comparison set

1,416Raw benchmark valueCI 1,396 - 1,436

Code Arena · Webdev React

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #31 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,390
Percentile: 49.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: webdev-react. Source rank: #39. Votes: 6716. Organization: openai. License: Proprietary.

49.2% percentile inside its fair comparison set

1,390Raw benchmark valueCI 1,382 - 1,398

Code Migration

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #17 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 12.9%
Percentile: 27.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: code-migration; provider: OpenAI.

27.3% percentile inside its fair comparison set

12.9%Raw benchmark valueCI 5.8% - 20.1%

LiveCodeBench

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #10 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 86.6%
Percentile: 90%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: lcb; provider: OpenAI.

90% percentile inside its fair comparison set

86.6%Raw benchmark valueCI 84.7% - 88.5%

ProgramBench

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #8 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 0%
Percentile: 70%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: programbench; provider: OpenAI.

70% percentile inside its fair comparison set

0%Raw benchmark valueCI 0% - 0%

SWE-bench Verified

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #27 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 73%
Percentile: 51.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: swebench; provider: OpenAI.

51.9% percentile inside its fair comparison set

73%Raw benchmark valueCI 69.1% - 76.9%

Terminal-Bench 2.1

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #14 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 54.7%
Percentile: 51.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: terminal-bench-2-1; provider: OpenAI.

51.9% percentile inside its fair comparison set

54.7%Raw benchmark valueCI 45.8% - 63.6%

Vibe Code Bench v1.1

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #13 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 48%
Percentile: 75.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: vibe-code; provider: OpenAI.

75.5% percentile inside its fair comparison set

48%Raw benchmark valueCI 37% - 59%

Text Arena · Coding

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #39 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,496
Percentile: 88.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: coding. Source rank: #49. Votes: 10646. Organization: openai. License: Proprietary.

88.1% percentile inside its fair comparison set

1,496Raw benchmark valueCI 1,489 - 1,502

Text Arena · Coding · No Style Control

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #65 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,438
Percentile: 80%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: coding. Source rank: #80. Votes: 10646. Organization: openai. License: Proprietary.

80% percentile inside its fair comparison set

1,438Raw benchmark valueCI 1,431 - 1,445

IOI

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #28 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 6.8%
Percentile: 38.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: ioi; provider: OpenAI.

38.6% percentile inside its fair comparison set

6.8%Raw benchmark valueCI 0.3% - 13.2%

Agentic coding

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #92 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 17%
Percentile: 15.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Agentic Coding. Tasks scored: 3.

15.7% percentile inside its fair comparison set

17%Raw benchmark value

Coding

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #82 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 68.2%
Percentile: 25%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Coding. Tasks scored: 2.

25% percentile inside its fair comparison set

68.2%Raw benchmark value

JavaScript

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #84 · Source label: gpt-5.4-mini-medium

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 21.1%
Percentile: 23.1%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: javascript. Category: Agentic Coding.

23.1% percentile inside its fair comparison set

21.1%Raw benchmark value

TypeScript

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #91 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 10%
Percentile: 16.8%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: typescript. Category: Agentic Coding.

16.8% percentile inside its fair comparison set

10%Raw benchmark value

Python

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #98 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 15%
Percentile: 10.2%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: python. Category: Agentic Coding.

10.2% percentile inside its fair comparison set

15%Raw benchmark value

Coding generation

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #89 · Source label: gpt-5.4-mini-xhigh

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 67.7%
Percentile: 18.5%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: code_generation. Category: Coding.

18.5% percentile inside its fair comparison set

67.7%Raw benchmark value

Coding completion

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #88 · Source label: gpt-5.4-mini-medium

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 64.4%
Percentile: 19.4%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: code_completion. Category: Coding.

19.4% percentile inside its fair comparison set

64.4%Raw benchmark value

Terminal-Bench 2.0

TERMINAL-BENCH · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #20 · Source label: gpt-5-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Terminal-Bench
Raw value: 31.9%
Percentile: 36.7%
Last updated: archived
Eligibility: headline eligible

Parsed from the public Terminal-Bench 2.0 verified leaderboard. Collapse policy: highest verified score per canonical model. Selected agent: Codex CLI (0.53.0). Display model: GPT-5-Mini. Integration method: API. Agent URL: https://developers.openai.com/codex/cli/. Reported stderr: 1.530 percentage points.

36.7% percentile inside its fair comparison set

31.9%Raw benchmark value

Reasoning / math / science21 benchmarks35.8%

Humanity's Last Exam

AA · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #219 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 5%
Percentile: 41.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `hle`.

41.1% percentile inside its fair comparison set

5%Raw benchmark value

GPQA

AA · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #197 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 60.6%
Percentile: 47.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `gpqa`.

47.6% percentile inside its fair comparison set

60.6%Raw benchmark value

CritPt

AA · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #106 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 0%
Percentile: 65.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `critpt`.

65.2% percentile inside its fair comparison set

0%Raw benchmark value

ProofBench

VALS-AI · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #15 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 21%
Percentile: 60%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: proof_bench; provider: OpenAI.

60% percentile inside its fair comparison set

21%Raw benchmark valueCI 13% - 29%

GPQA Diamond

VALS-AI · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #36 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 83.1%
Percentile: 60.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: gpqa; provider: OpenAI.

60.7% percentile inside its fair comparison set

83.1%Raw benchmark valueCI 78.3% - 87.9%

MMLU Pro

VALS-AI · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #38 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 84.6%
Percentile: 58.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: mmlu_pro; provider: OpenAI.

58.4% percentile inside its fair comparison set

84.6%Raw benchmark valueCI 83.9% - 85.3%

Text Arena · Math

AR · Reasoning / math / science · Human

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #43 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,441
Percentile: 86.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: math. Source rank: #54. Votes: 2233. Organization: openai. License: Proprietary.

86.6% percentile inside its fair comparison set

1,441Raw benchmark valueCI 1,428 - 1,454

Text Arena · Math · No Style Control

AR · Reasoning / math / science · Human

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #69 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,421
Percentile: 78.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: math. Source rank: #83. Votes: 2233. Organization: openai. License: Proprietary.

78.3% percentile inside its fair comparison set

1,421Raw benchmark valueCI 1,408 - 1,434

TutorBench

SL · Reasoning / math / science · Rubric

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 55.3%
Percentile: 80%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Manually verified from the official Scale Labs TutorBench leaderboard. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

80% percentile inside its fair comparison set

55.3%Raw benchmark value

MultiNRC

SL · Reasoning / math / science · Rubric

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #6 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 52.1%
Percentile: 60%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

60% percentile inside its fair comparison set

52.1%Raw benchmark value

EnigmaEval

SL · Reasoning / math / science · Rubric

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #6 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 64%
Percentile: 80%
Last updated: aging
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

80% percentile inside its fair comparison set

64%Raw benchmark value

Mathematics

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #107 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 37%
Percentile: 1.9%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Mathematics. Tasks scored: 4.

1.9% percentile inside its fair comparison set

37%Raw benchmark value

Reasoning

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #107 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 21.9%
Percentile: 1.9%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Reasoning. Tasks scored: 4.

1.9% percentile inside its fair comparison set

21.9%Raw benchmark value

AMPS Hard

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 68%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: AMPS_Hard. Category: Mathematics.

4.6% percentile inside its fair comparison set

68%Raw benchmark value

Integrals with game

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #104 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 6%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: integrals_with_game. Category: Mathematics.

4.6% percentile inside its fair comparison set

6%Raw benchmark value

Math competition

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #109 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 18.6%
Percentile: 0%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: math_comp. Category: Mathematics.

0% percentile inside its fair comparison set

18.6%Raw benchmark value

Olympiad

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #103 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 51.3%
Percentile: 5.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: olympiad. Category: Mathematics.

5.6% percentile inside its fair comparison set

51.3%Raw benchmark value

Theory of mind

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #105 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 28.8%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: theory_of_mind. Category: Reasoning.

3.7% percentile inside its fair comparison set

28.8%Raw benchmark value

Zebra puzzle

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #103 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 6.7%
Percentile: 4.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: zebra_puzzle. Category: Reasoning.

4.7% percentile inside its fair comparison set

6.7%Raw benchmark value

Spatial

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #105 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 46%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: spatial. Category: Reasoning.

3.7% percentile inside its fair comparison set

46%Raw benchmark value

Logic with navigation

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #106 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 6%
Percentile: 2.8%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: logic_with_navigation. Category: Reasoning.

2.8% percentile inside its fair comparison set

6%Raw benchmark value

Professional reasoning35 benchmarks64.4%

GDPval-AA

AA · Professional reasoning · Rubric

Agentic performance on economically valuable work tasks.

Rank #35 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 755
Percentile: 26.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `gdpvalBreakdown.elo`.

26.1% percentile inside its fair comparison set

755Raw benchmark value

APEX-Agents-AA

AA · Professional reasoning · Objective

Long-horizon agentic task completion.

Rank #7 · Source label: GPT-5.4 mini (xhigh)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 28.2%
Percentile: 75%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `apexAgents`.

75% percentile inside its fair comparison set

28.2%Raw benchmark value

Legal Research Bench

VALS-AI · Professional reasoning · Objective

Applied legal research tasks.

Rank #13 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 12.5%
Percentile: 0%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: legal_research; provider: OpenAI.

0% percentile inside its fair comparison set

12.5%Raw benchmark valueCI 8% - 17%

Vals Index

VALS-AI · Professional reasoning · Combined

Weighted model performance across economically relevant Vals tasks.

Rank #14 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 52
Percentile: 50%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: vals_index; provider: OpenAI.

50% percentile inside its fair comparison set

52Raw benchmark valueCI 48 - 56

LegalBench

VALS-AI · Professional reasoning · Objective

Academic legal reasoning tasks.

Rank #47 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 81.8%
Percentile: 48.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: legal_bench; provider: OpenAI.

48.9% percentile inside its fair comparison set

81.8%Raw benchmark valueCI 80.9% - 82.7%

Finance Agent v2

VALS-AI · Professional reasoning · Objective

Core financial analyst tasks for agentic models.

Rank #9 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 45.4%
Percentile: 68%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: fabv2; provider: OpenAI.

68% percentile inside its fair comparison set

45.4%Raw benchmark valueCI 44.5% - 46.2%

TaxEval v2

VALS-AI · Professional reasoning · Objective

Answer quality on tax questions and responses.

Rank #11 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 75.2%
Percentile: 89%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: tax_eval_v2; provider: OpenAI.

89% percentile inside its fair comparison set

75.2%Raw benchmark valueCI 73.6% - 76.9%

MedCode

VALS-AI · Professional reasoning · Objective

Medical billing support and coding tasks.

Rank #22 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 43%
Percentile: 58.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: medcode; provider: OpenAI.

58.8% percentile inside its fair comparison set

43%Raw benchmark valueCI 39% - 47.1%

MedScribe

VALS-AI · Professional reasoning · Objective

Administrative documentation support for doctors.

Rank #19 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 80.6%
Percentile: 64%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: medscribe; provider: OpenAI.

64% percentile inside its fair comparison set

80.6%Raw benchmark valueCI 76.8% - 84.3%

Text Arena · Expert

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena expert leaderboard.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 88.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: expert. Source rank: #39. Votes: 3590. Organization: openai. License: Proprietary.

88.7% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,471 - 1,492

Text Arena · Industry Business And Management And Financial Operations

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_business_and_management_and_financial_operations leaderboard.

Rank #22 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,460
Percentile: 93.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_business_and_management_and_financial_operations. Source rank: #28. Votes: 7999. Organization: openai. License: Proprietary.

93.4% percentile inside its fair comparison set

1,460Raw benchmark valueCI 1,452 - 1,468

Text Arena · Industry Entertainment And Sports And Media

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_entertainment_and_sports_and_media leaderboard.

Rank #47 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,408
Percentile: 85.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_entertainment_and_sports_and_media. Source rank: #59. Votes: 8212. Organization: openai. License: Proprietary.

85.8% percentile inside its fair comparison set

1,408Raw benchmark valueCI 1,400 - 1,415

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_legal_and_government leaderboard.

Rank #41 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,457
Percentile: 86.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_legal_and_government. Source rank: #54. Votes: 3015. Organization: openai. License: Proprietary.

86.6% percentile inside its fair comparison set

1,457Raw benchmark valueCI 1,445 - 1,469

Text Arena · Industry Life And Physical And Social Science

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_life_and_physical_and_social_science leaderboard.

Rank #41 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,465
Percentile: 87.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_life_and_physical_and_social_science. Source rank: #52. Votes: 6537. Organization: openai. License: Proprietary.

87.6% percentile inside its fair comparison set

1,465Raw benchmark valueCI 1,457 - 1,474

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_mathematical leaderboard.

Rank #37 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,456
Percentile: 88.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_mathematical. Source rank: #45. Votes: 2140. Organization: openai. License: Proprietary.

88.3% percentile inside its fair comparison set

1,456Raw benchmark valueCI 1,443 - 1,470

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_medicine_and_healthcare leaderboard.

Rank #55 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,455
Percentile: 81.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_medicine_and_healthcare. Source rank: #71. Votes: 2894. Organization: openai. License: Proprietary.

81.7% percentile inside its fair comparison set

1,455Raw benchmark valueCI 1,443 - 1,467

Text Arena · Industry Software And It Services

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_software_and_it_services leaderboard.

Rank #38 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 88.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_software_and_it_services. Source rank: #49. Votes: 15521. Organization: openai. License: Proprietary.

88.6% percentile inside its fair comparison set

1,487Raw benchmark valueCI 1,481 - 1,493

Text Arena · Industry Writing And Literature And Language

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_writing_and_literature_and_language leaderboard.

Rank #44 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,423
Percentile: 86.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_writing_and_literature_and_language. Source rank: #57. Votes: 9541. Organization: openai. License: Proprietary.

86.7% percentile inside its fair comparison set

1,423Raw benchmark valueCI 1,416 - 1,430

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena expert leaderboard.

Rank #57 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,436
Percentile: 79.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: expert. Source rank: #68. Votes: 3590. Organization: openai. License: Proprietary.

79.6% percentile inside its fair comparison set

1,436Raw benchmark valueCI 1,425 - 1,446

Text Arena · Industry Business And Management And Financial Operations · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_business_and_management_and_financial_operations leaderboard.

Rank #59 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,417
Percentile: 81.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_business_and_management_and_financial_operations. Source rank: #71. Votes: 7999. Organization: openai. License: Proprietary.

81.8% percentile inside its fair comparison set

1,417Raw benchmark valueCI 1,410 - 1,425

Text Arena · Industry Entertainment And Sports And Media · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_entertainment_and_sports_and_media leaderboard.

Rank #82 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,372
Percentile: 74.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_entertainment_and_sports_and_media. Source rank: #100. Votes: 8212. Organization: openai. License: Proprietary.

74.9% percentile inside its fair comparison set

1,372Raw benchmark valueCI 1,364 - 1,379

Text Arena · Industry Legal And Government · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_legal_and_government leaderboard.

Rank #83 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,415
Percentile: 72.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_legal_and_government. Source rank: #101. Votes: 3015. Organization: openai. License: Proprietary.

72.5% percentile inside its fair comparison set

1,415Raw benchmark valueCI 1,403 - 1,427

Text Arena · Industry Life And Physical And Social Science · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_life_and_physical_and_social_science leaderboard.

Rank #91 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,417
Percentile: 72.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_life_and_physical_and_social_science. Source rank: #107. Votes: 6537. Organization: openai. License: Proprietary.

72.1% percentile inside its fair comparison set

1,417Raw benchmark valueCI 1,409 - 1,425

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_mathematical leaderboard.

Rank #58 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 81.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_mathematical. Source rank: #68. Votes: 2140. Organization: openai. License: Proprietary.

81.5% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,419 - 1,446

Text Arena · Industry Medicine And Healthcare · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_medicine_and_healthcare leaderboard.

Rank #101 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,397
Percentile: 66.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_medicine_and_healthcare. Source rank: #122. Votes: 2894. Organization: openai. License: Proprietary.

66.1% percentile inside its fair comparison set

1,397Raw benchmark valueCI 1,385 - 1,409

Text Arena · Industry Software And It Services · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_software_and_it_services leaderboard.

Rank #71 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,436
Percentile: 78.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_software_and_it_services. Source rank: #86. Votes: 15521. Organization: openai. License: Proprietary.

78.5% percentile inside its fair comparison set

1,436Raw benchmark valueCI 1,430 - 1,442

Text Arena · Industry Writing And Literature And Language · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_writing_and_literature_and_language leaderboard.

Rank #71 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,392
Percentile: 78.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_writing_and_literature_and_language. Source rank: #86. Votes: 9541. Organization: openai. License: Proprietary.

78.4% percentile inside its fair comparison set

1,392Raw benchmark valueCI 1,385 - 1,399

SAGE

VALS-AI · Professional reasoning · Objective

Student Assessment with Generative Evaluation.

Rank #9 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 50.8%
Percentile: 82.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: sage; provider: OpenAI.

82.2% percentile inside its fair comparison set

50.8%Raw benchmark valueCI 44.1% - 57.5%

PRBench Legal

SL · Professional reasoning · Rubric

Applied legal reasoning on professional-domain tasks.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 49%
Percentile: 83.3%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

83.3% percentile inside its fair comparison set

49%Raw benchmark value

Data analysis

LB · Professional reasoning · Objective

Structured data manipulation and table reasoning accuracy.

Rank #101 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 41.5%
Percentile: 7.4%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Data Analysis. Tasks scored: 3.

7.4% percentile inside its fair comparison set

41.5%Raw benchmark value

Overall

LB · Professional reasoning · Objective

Average objective performance across LiveBench's current public category mix.

Rank #103 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 37%
Percentile: 5.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category averages included: 7.

5.6% percentile inside its fair comparison set

37%Raw benchmark value

Consecutive events

LB · Professional reasoning · Objective

Objective consecutive events score in LiveBench.

Rank #103 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 0.5%
Percentile: 5.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: consecutive_events. Category: Data Analysis.

5.6% percentile inside its fair comparison set

0.5%Raw benchmark value

Table join

LB · Professional reasoning · Objective

Objective table join score in LiveBench.

Rank #99 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 29.9%
Percentile: 9.3%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: tablejoin. Category: Data Analysis.

9.3% percentile inside its fair comparison set

29.9%Raw benchmark value

Table reformat

LB · Professional reasoning · Objective

Objective table reformat score in LiveBench.

Rank #93 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 94.1%
Percentile: 14.8%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: tablereformat. Category: Data Analysis.

14.8% percentile inside its fair comparison set

94.1%Raw benchmark value

Poker Agent

VALS-AI · Professional reasoning · Objective

Agent profit in poker-style strategic play.

Rank #4 · Source label: openai/gpt-5-2025-08-07

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 1,103.2 score
Percentile: 93.8%
Last updated: archived
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Parsed from Vals AI BenchmarkView overall scores. Vals slug: poker_agent; provider: unknown. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

93.8% percentile inside its fair comparison set

1,103.2 scoreRaw benchmark valueCI 1,103.2 score - 1,103.2 score

Search / tool use3 benchmarks31.7%

Tau2-Bench Telecom

AA · Search / tool use · Objective

It matters when the model must browse, call tools, and recover useful answers from external systems.

Rank #212 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.4%
Percentile: 31.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `tau2`.

31.7% percentile inside its fair comparison set

23.4%Raw benchmark value

Search Arena

AR · Search / tool use · Human

It matters when the model must browse, call tools, and recover useful answers from external systems.

Rank #19 · Source label: gpt-5-search

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,180
Percentile: 43.3%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Parsed from Arena leaderboard dataset row `gpt-5-search`. Category: overall. Source rank: #19. Votes: 20928. Organization: openai. License: Proprietary. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

43.3% percentile inside its fair comparison set

1,180Raw benchmark valueCI 1,173 - 1,188

Search Arena · No Style Control

AR · Search / tool use · Human

It matters when the model must browse, call tools, and recover useful answers from external systems.

Rank #26 · Source label: gpt-5-search

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,132
Percentile: 20%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Parsed from Arena leaderboard dataset row `gpt-5-search`. Category: overall. Source rank: #27. Votes: 20928. Organization: openai. License: Proprietary. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

20% percentile inside its fair comparison set

1,132Raw benchmark valueCI 1,127 - 1,138

Long context2 benchmarks50%

Long Context Reasoning

AA · Long context · Objective

It checks whether long-context claims survive contact with retrieval, memory, or long-document tasks.

Rank #155 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 30.3%
Percentile: 51.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `lcr`.

51.1% percentile inside its fair comparison set

30.3%Raw benchmark value

CorpFin v2

VALS-AI · Long context · Objective

It checks whether long-context claims survive contact with retrieval, memory, or long-document tasks.

Rank #46 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 60.9%
Percentile: 48.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: corp_fin_v2; provider: OpenAI.

48.9% percentile inside its fair comparison set

60.9%Raw benchmark valueCI 59% - 62.8%

Vision understanding24 benchmarks56.9%

MMMU-Pro

AA · Vision understanding · Objective

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #81 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 58.4%
Percentile: 40.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `mmmuPro`.

40.7% percentile inside its fair comparison set

58.4%Raw benchmark value

Vision Arena

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #17 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,256
Percentile: 85.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #23. Votes: 15692. Organization: openai. License: Proprietary.

85.3% percentile inside its fair comparison set

1,256Raw benchmark valueCI 1,249 - 1,263

Vision Arena · Captioning

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,165
Percentile: 26.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: captioning. Source rank: #21. Votes: 302. Organization: openai. License: Proprietary.

26.9% percentile inside its fair comparison set

1,165Raw benchmark valueCI 1,125 - 1,206

Vision Arena · Creative Writing Vision

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #29 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,233
Percentile: 49.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing_vision. Source rank: #36. Votes: 897. Organization: openai. License: Proprietary.

49.1% percentile inside its fair comparison set

1,233Raw benchmark valueCI 1,212 - 1,254

Vision Arena · Diagram

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #17 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,281
Percentile: 77.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: diagram. Source rank: #23. Votes: 4259. Organization: openai. License: Proprietary.

77.1% percentile inside its fair comparison set

1,281Raw benchmark valueCI 1,270 - 1,292

Vision Arena · English

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,252
Percentile: 82.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #25. Votes: 6782. Organization: openai. License: Proprietary.

82.6% percentile inside its fair comparison set

1,252Raw benchmark valueCI 1,242 - 1,262

Vision Arena · Entity Recognition

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,205
Percentile: 40.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: entity_recognition. Source rank: #19. Votes: 283. Organization: openai. License: Proprietary.

40.6% percentile inside its fair comparison set

1,205Raw benchmark valueCI 1,165 - 1,245

Vision Arena · Homework

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #16 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,290
Percentile: 77.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: homework. Source rank: #21. Votes: 2433. Organization: openai. License: Proprietary.

77.9% percentile inside its fair comparison set

1,290Raw benchmark valueCI 1,277 - 1,303

Vision Arena · Humor

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #22 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,237
Percentile: 57.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: humor. Source rank: #28. Votes: 498. Organization: openai. License: Proprietary.

57.1% percentile inside its fair comparison set

1,237Raw benchmark valueCI 1,209 - 1,265

Vision Arena · Ocr

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #18 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,267
Percentile: 75.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: ocr. Source rank: #24. Votes: 11142. Organization: openai. License: Proprietary.

75.7% percentile inside its fair comparison set

1,267Raw benchmark valueCI 1,259 - 1,274

Vision Arena · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,250
Percentile: 75.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #35. Votes: 15692. Organization: openai. License: Proprietary.

75.2% percentile inside its fair comparison set

1,250Raw benchmark valueCI 1,242 - 1,257

Vision Arena · Captioning · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #16 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,213
Percentile: 42.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: captioning. Source rank: #15. Votes: 302. Organization: openai. License: Proprietary.

42.3% percentile inside its fair comparison set

1,213Raw benchmark valueCI 1,179 - 1,247

Vision Arena · Creative Writing Vision · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #36 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,227
Percentile: 36.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing_vision. Source rank: #44. Votes: 897. Organization: openai. License: Proprietary.

36.4% percentile inside its fair comparison set

1,227Raw benchmark valueCI 1,206 - 1,248

Vision Arena · Diagram · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,263
Percentile: 61.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: diagram. Source rank: #35. Votes: 4259. Organization: openai. License: Proprietary.

61.4% percentile inside its fair comparison set

1,263Raw benchmark valueCI 1,252 - 1,273

Vision Arena · English · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,246
Percentile: 71.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #39. Votes: 6782. Organization: openai. License: Proprietary.

71.6% percentile inside its fair comparison set

1,246Raw benchmark valueCI 1,236 - 1,256

Vision Arena · Entity Recognition · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,225
Percentile: 40.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: entity_recognition. Source rank: #19. Votes: 283. Organization: openai. License: Proprietary.

40.6% percentile inside its fair comparison set

1,225Raw benchmark valueCI 1,191 - 1,259

Vision Arena · Homework · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #21 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,285
Percentile: 70.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: homework. Source rank: #27. Votes: 2433. Organization: openai. License: Proprietary.

70.6% percentile inside its fair comparison set

1,285Raw benchmark valueCI 1,272 - 1,298

Vision Arena · Humor · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #27 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,239
Percentile: 46.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: humor. Source rank: #33. Votes: 498. Organization: openai. License: Proprietary.

46.9% percentile inside its fair comparison set

1,239Raw benchmark valueCI 1,211 - 1,267

Vision Arena · Ocr · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,256
Percentile: 61.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: ocr. Source rank: #36. Votes: 11142. Organization: openai. License: Proprietary.

61.4% percentile inside its fair comparison set

1,256Raw benchmark valueCI 1,249 - 1,264

MMMU Pro

VALS-AI · Vision understanding · Objective

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #31 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 79.2%
Percentile: 48.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: mmmu; provider: OpenAI.

48.3% percentile inside its fair comparison set

79.2%Raw benchmark valueCI 77.3% - 81.2%

VTB

SL · Vision understanding · Rubric

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #6 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 17%
Percentile: 63.6%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

63.6% percentile inside its fair comparison set

17%Raw benchmark value

VISTA

SL · Vision understanding · Rubric

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 79%
Percentile: 92.9%
Last updated: aging
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

92.9% percentile inside its fair comparison set

79%Raw benchmark value

Vision Arena · Creative Writing

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,152
Percentile: 15.6%
Last updated: archived
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: creative_writing. Source rank: #29. Votes: 1108. Organization: openai. License: Proprietary.

15.6% percentile inside its fair comparison set

1,152Raw benchmark valueCI 1,132 - 1,173

Vision Arena · Creative Writing · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #25 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,181
Percentile: 25%
Last updated: archived
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: creative_writing. Source rank: #26. Votes: 1108. Organization: openai. License: Proprietary.

25% percentile inside its fair comparison set

1,181Raw benchmark valueCI 1,163 - 1,199

Document understanding3 benchmarks70.7%

Vals Multimodal Index

VALS-AI · Document understanding · Combined

It matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.

Rank #10 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 54
Percentile: 52.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: vals_multimodal_index; provider: OpenAI.

52.6% percentile inside its fair comparison set

54Raw benchmark valueCI 51 - 58

MortgageTax

VALS-AI · Document understanding · Objective

It matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.

Rank #21 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 66.9%
Percentile: 66.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: mortgage_tax; provider: OpenAI.

66.7% percentile inside its fair comparison set

66.9%Raw benchmark valueCI 65.1% - 68.7%

Multimodal mix

OC · Document understanding · Objective

It matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: OpenCompass
Raw value: 75.4%
Percentile: 92.9%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

92.9% percentile inside its fair comparison set

75.4%Raw benchmark value

Safety1 benchmark61.5%

MASK

SL · Safety · Rubric

Whether a model stays honest instead of covertly optimizing against the user.

Rank #7 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 79.3%
Percentile: 61.5%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

61.5% percentile inside its fair comparison set

79.3%Raw benchmark value

Embeddings / retrieval1 benchmark100%

Retrieval

MTEB · Embeddings / retrieval · Retrieval

It is one of the few direct signals for retrieval stacks, where embedding quality matters more than chat style.

Rank #3 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: MTEB
Raw value: 58.8 ndcg
Percentile: 100%
Last updated: aging
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Embedding endpoint score. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

100% percentile inside its fair comparison set

58.8 ndcgRaw benchmark value

Multilingual16 benchmarks77.4%

Text Arena · Chinese

AR · Multilingual · Human

Observed user preference in Arena's Text Arena chinese leaderboard.

Rank #50 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 83.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #61. Votes: 2151. Organization: openai. License: Proprietary.

83.4% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,465 - 1,493

Text Arena · French

AR · Multilingual · Human

Observed user preference in Arena's Text Arena french leaderboard.

Rank #41 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,465
Percentile: 81.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: french. Source rank: #52. Votes: 1303. Organization: openai. License: Proprietary.

81.5% percentile inside its fair comparison set

1,465Raw benchmark valueCI 1,446 - 1,484

Text Arena · German

AR · Multilingual · Human

Observed user preference in Arena's Text Arena german leaderboard.

Rank #42 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,437
Percentile: 82.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: german. Source rank: #57. Votes: 537. Organization: openai. License: Proprietary.

82.7% percentile inside its fair comparison set

1,437Raw benchmark valueCI 1,411 - 1,463

Text Arena · Japanese

AR · Multilingual · Human

Observed user preference in Arena's Text Arena japanese leaderboard.

Rank #31 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,415
Percentile: 85.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: japanese. Source rank: #43. Votes: 281. Organization: openai. License: Proprietary.

85.2% percentile inside its fair comparison set

1,415Raw benchmark valueCI 1,377 - 1,452

Text Arena · Korean

AR · Multilingual · Human

Observed user preference in Arena's Text Arena korean leaderboard.

Rank #51 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,384
Percentile: 76%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: korean. Source rank: #65. Votes: 607. Organization: openai. License: Proprietary.

76% percentile inside its fair comparison set

1,384Raw benchmark valueCI 1,358 - 1,409

Text Arena · Russian

AR · Multilingual · Human

Observed user preference in Arena's Text Arena russian leaderboard.

Rank #34 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,453
Percentile: 88.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: russian. Source rank: #43. Votes: 4221. Organization: openai. License: Proprietary.

88.6% percentile inside its fair comparison set

1,453Raw benchmark valueCI 1,443 - 1,462

Text Arena · Spanish

AR · Multilingual · Human

Observed user preference in Arena's Text Arena spanish leaderboard.

Rank #39 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,443
Percentile: 82.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: spanish. Source rank: #49. Votes: 1135. Organization: openai. License: Proprietary.

82.2% percentile inside its fair comparison set

1,443Raw benchmark valueCI 1,423 - 1,462

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena chinese leaderboard.

Rank #77 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,451
Percentile: 74.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #93. Votes: 2151. Organization: openai. License: Proprietary.

74.2% percentile inside its fair comparison set

1,451Raw benchmark valueCI 1,437 - 1,465

Text Arena · French · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena french leaderboard.

Rank #63 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,431
Percentile: 71.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: french. Source rank: #78. Votes: 1303. Organization: openai. License: Proprietary.

71.3% percentile inside its fair comparison set

1,431Raw benchmark valueCI 1,412 - 1,449

Text Arena · German · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena german leaderboard.

Rank #70 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,403
Percentile: 70.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: german. Source rank: #87. Votes: 537. Organization: openai. License: Proprietary.

70.9% percentile inside its fair comparison set

1,403Raw benchmark valueCI 1,377 - 1,428

Text Arena · Japanese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena japanese leaderboard.

Rank #46 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,383
Percentile: 77.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: japanese. Source rank: #59. Votes: 281. Organization: openai. License: Proprietary.

77.8% percentile inside its fair comparison set

1,383Raw benchmark valueCI 1,345 - 1,421

Text Arena · Korean · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena korean leaderboard.

Rank #74 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,358
Percentile: 64.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: korean. Source rank: #90. Votes: 607. Organization: openai. License: Proprietary.

64.9% percentile inside its fair comparison set

1,358Raw benchmark valueCI 1,333 - 1,384

Text Arena · Russian · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena russian leaderboard.

Rank #55 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,419
Percentile: 81.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: russian. Source rank: #69. Votes: 4221. Organization: openai. License: Proprietary.

81.3% percentile inside its fair comparison set

1,419Raw benchmark valueCI 1,409 - 1,429

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena spanish leaderboard.

Rank #66 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,413
Percentile: 69.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: spanish. Source rank: #80. Votes: 1135. Organization: openai. License: Proprietary.

69.6% percentile inside its fair comparison set

1,413Raw benchmark valueCI 1,393 - 1,432

Vision Arena · Chinese

AR · Multilingual · Human

Observed user preference in Arena's Vision Arena chinese leaderboard.

Rank #17 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,301
Percentile: 79.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #22. Votes: 899. Organization: openai. License: Proprietary.

79.2% percentile inside its fair comparison set

1,301Raw benchmark valueCI 1,276 - 1,326

Vision Arena · Chinese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Vision Arena chinese leaderboard.

Rank #24 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,302
Percentile: 70.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #29. Votes: 899. Organization: openai. License: Proprietary.

70.1% percentile inside its fair comparison set

1,302Raw benchmark valueCI 1,278 - 1,327

Source links and registry checks

official

OpenAI models docs

Jun 20, 2026

source →

official

Terminal-Bench

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Artificial Analysis

Jun 20, 2026

source →

official

LiveBench

Jun 20, 2026

source →

Model profile · OpenAI

GPT-5.4 mini

Closed weightspremium · registry tag 2026 mini

Visible tradeoffs

Reads as visible tradeoffs across the resolved source data.

Visible coverage: 63%
Verified coverage: 55.5%
Spread: 90%
Last verified: Jun 20, 2026

42%bench fit

textcodevisiondocumentsearch19 aliases46 official source links

Open compare

Data version

Current snapshot.

Data version Jun 20, 2026Model list checked9 providers · 1081 tracked modelsPage refreshed Jul 5, 2026

The registry snapshot and page stamp are shown so a stale deploy is visible at a glance.

Source-linked scores by benchmark

Each row keeps the benchmark source, source type, raw metric, and percentile inside its fair comparison set.

Visible tradeoffsThis model currently reads as visible tradeoffs across the resolved source data.

Chat / text38 benchmarks48.5%

Intelligence Index

AA · Chat / text · Combined

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #171 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 14
Percentile: 57%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `intelligenceIndex`.

57% percentile inside its fair comparison set

14Raw benchmark value

AA-Omniscience accuracy

AA · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #136 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 18.2%
Percentile: 54.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceAccuracy`.

54.7% percentile inside its fair comparison set

18.2%Raw benchmark value

AA-Omniscience non-hallucination

AA · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #276 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 4.7%
Percentile: 7.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceNonHallucination`.

7.7% percentile inside its fair comparison set

4.7%Raw benchmark value

IFBench

AA · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #182 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 38.8%
Percentile: 42.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `ifbench`.

42.5% percentile inside its fair comparison set

38.8%Raw benchmark value

Blended price

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #199 · Source label: GPT-5.4 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $1.7 /1M tokens
Percentile: 28.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mBlended0To3To1`.

28.3% percentile inside its fair comparison set

$1.7 /1M tokensRaw benchmark value

Input price

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #189 · Source label: GPT-5.4 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $0.8 /1M input tokens
Percentile: 31.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mInputTokens`.

31.9% percentile inside its fair comparison set

$0.8 /1M input tokensRaw benchmark value

Output price

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #206 · Source label: GPT-5.4 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $4.5 /1M output tokens
Percentile: 25.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mOutputTokens`.

25.7% percentile inside its fair comparison set

$4.5 /1M output tokensRaw benchmark value

Output Speed

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #105 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 94.6 tokens/s
Percentile: 50.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianOutputTokensPerSecond`.

50.5% percentile inside its fair comparison set

94.6 tokens/sRaw benchmark value

Time to first token

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #207 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 104.11s
Percentile: 1.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstTokenSeconds`.

1.9% percentile inside its fair comparison set

104.11sRaw benchmark value

Time to first answer token

AA · Chat / text · Speed / cost

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #206 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 104.11s
Percentile: 2.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstAnswerTokenSeconds`.

2.4% percentile inside its fair comparison set

104.11sRaw benchmark value

Openness Index

AA · Chat / text · Combined

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #183 · Source label: GPT-5 mini (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 6
Percentile: 7.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `opennessBreakdown.opennessIndex`.

7.5% percentile inside its fair comparison set

6Raw benchmark value

Text Arena

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #39 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,448
Percentile: 88.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #50. Votes: 39525. Organization: openai. License: Proprietary.

88.3% percentile inside its fair comparison set

1,448Raw benchmark valueCI 1,444 - 1,452

Text Arena · Creative Writing

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #61 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,399
Percentile: 81.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing. Source rank: #75. Votes: 6391. Organization: openai. License: Proprietary.

81.4% percentile inside its fair comparison set

1,399Raw benchmark valueCI 1,390 - 1,407

Text Arena · English

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #42 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,452
Percentile: 87.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #54. Votes: 18792. Organization: openai. License: Proprietary.

87.4% percentile inside its fair comparison set

1,452Raw benchmark valueCI 1,446 - 1,457

Text Arena · Exclude Ties

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #37 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,450
Percentile: 88.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: exclude_ties. Source rank: #48. Votes: 30409. Organization: openai. License: Proprietary.

88.9% percentile inside its fair comparison set

1,450Raw benchmark valueCI 1,445 - 1,456

Text Arena · Hard Prompts

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #37 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,471
Percentile: 88.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts. Source rank: #49. Votes: 25673. Organization: openai. License: Proprietary.

88.9% percentile inside its fair comparison set

1,471Raw benchmark valueCI 1,466 - 1,476

Text Arena · Hard Prompts English

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #43 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 87%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts_english. Source rank: #55. Votes: 12804. Organization: openai. License: Proprietary.

87% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,462 - 1,475

Text Arena · Instruction Following

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #44 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 86.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: instruction_following. Source rank: #55. Votes: 13187. Organization: openai. License: Proprietary.

86.8% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,426 - 1,439

Text Arena · Longer Query

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #48 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,446
Percentile: 84.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: longer_query. Source rank: #61. Votes: 16352. Organization: openai. License: Proprietary.

84.5% percentile inside its fair comparison set

1,446Raw benchmark valueCI 1,440 - 1,453

Text Arena · Multi Turn

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #25 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,472
Percentile: 92.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: multi_turn. Source rank: #34. Votes: 7385. Organization: openai. License: Proprietary.

92.6% percentile inside its fair comparison set

1,472Raw benchmark valueCI 1,464 - 1,480

Text Arena · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #81 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,412
Percentile: 75.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #97. Votes: 39525. Organization: openai. License: Proprietary.

75.4% percentile inside its fair comparison set

1,412Raw benchmark valueCI 1,408 - 1,416

Text Arena · Creative Writing · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #85 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,367
Percentile: 74%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing. Source rank: #105. Votes: 6391. Organization: openai. License: Proprietary.

74% percentile inside its fair comparison set

1,367Raw benchmark valueCI 1,359 - 1,375

Text Arena · English · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #93 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,413
Percentile: 71.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #109. Votes: 18792. Organization: openai. License: Proprietary.

71.7% percentile inside its fair comparison set

1,413Raw benchmark valueCI 1,407 - 1,418

Text Arena · Exclude Ties · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #81 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 75.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: exclude_ties. Source rank: #96. Votes: 30409. Organization: openai. License: Proprietary.

75.4% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,393 - 1,404

Text Arena · Hard Prompts · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #69 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,426
Percentile: 79.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts. Source rank: #84. Votes: 25673. Organization: openai. License: Proprietary.

79.1% percentile inside its fair comparison set

1,426Raw benchmark valueCI 1,421 - 1,431

Text Arena · Hard Prompts English · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #82 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,422
Percentile: 75%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: hard_prompts_english. Source rank: #98. Votes: 12804. Organization: openai. License: Proprietary.

75% percentile inside its fair comparison set

1,422Raw benchmark valueCI 1,415 - 1,428

Text Arena · Instruction Following · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #66 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,403
Percentile: 80%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: instruction_following. Source rank: #80. Votes: 13187. Organization: openai. License: Proprietary.

80% percentile inside its fair comparison set

1,403Raw benchmark valueCI 1,397 - 1,409

Text Arena · Longer Query · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #80 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,407
Percentile: 74%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: longer_query. Source rank: #98. Votes: 16352. Organization: openai. License: Proprietary.

74% percentile inside its fair comparison set

1,407Raw benchmark valueCI 1,401 - 1,413

Text Arena · Multi Turn · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #48 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,434
Percentile: 85.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: multi_turn. Source rank: #60. Votes: 7385. Organization: openai. License: Proprietary.

85.4% percentile inside its fair comparison set

1,434Raw benchmark valueCI 1,427 - 1,442

Instruction following

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #102 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 18.9%
Percentile: 6.5%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: IF. Tasks scored: 4.

6.5% percentile inside its fair comparison set

18.9%Raw benchmark value

Language

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #105 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 41.8%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Language. Tasks scored: 3.

3.7% percentile inside its fair comparison set

41.8%Raw benchmark value

Paraphrase

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 14.6%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: paraphrase. Category: IF.

4.6% percentile inside its fair comparison set

14.6%Raw benchmark value

Simplify

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #92 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 22.9%
Percentile: 15.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: simplify. Category: IF.

15.7% percentile inside its fair comparison set

22.9%Raw benchmark value

Story generation

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #102 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 16.8%
Percentile: 6.5%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: story_generation. Category: IF.

6.5% percentile inside its fair comparison set

16.8%Raw benchmark value

Summarize

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 20.2%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: summarize. Category: IF.

4.6% percentile inside its fair comparison set

20.2%Raw benchmark value

Connections

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #100 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 48.5%
Percentile: 8.3%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: connections. Category: Language.

8.3% percentile inside its fair comparison set

48.5%Raw benchmark value

Plot unscrambling

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 25.9%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: plot_unscrambling. Category: Language.

4.6% percentile inside its fair comparison set

25.9%Raw benchmark value

Typos

LB · Chat / text · Objective

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 46%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: typos. Category: Language.

3.7% percentile inside its fair comparison set

46%Raw benchmark value

Coding25 benchmarks49.9%

Terminal-Bench Hard

AA · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #128 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 14.4%
Percentile: 57.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `terminalbenchHard`.

57.9% percentile inside its fair comparison set

14.4%Raw benchmark value

SciCode

AA · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #99 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 36.9%
Percentile: 73.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `scicode`.

73.4% percentile inside its fair comparison set

36.9%Raw benchmark value

Coding Index

AA · Coding · Combined

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #17 · Source label: GPT-5.4 mini (xhigh)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 56
Percentile: 78.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `codingIndex`.

78.7% percentile inside its fair comparison set

56Raw benchmark value

Agentic Index

AA · Coding · Combined

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #13 · Source label: GPT-5.4 mini (xhigh)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 30
Percentile: 73.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `agenticIndex`.

73.9% percentile inside its fair comparison set

30Raw benchmark value

Code Arena

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 57.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #40. Votes: 7713. Organization: openai. License: Proprietary.

57.5% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,391 - 1,406

WebDev Arena

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 57.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: webdev. Source rank: #40. Votes: 7713. Organization: openai. License: Proprietary.

57.5% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,391 - 1,406

Code Arena · Webdev Html

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #30 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,416
Percentile: 60.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: webdev-html. Source rank: #38. Votes: 960. Organization: openai. License: Proprietary.

60.3% percentile inside its fair comparison set

1,416Raw benchmark valueCI 1,396 - 1,436

Code Arena · Webdev React

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #31 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,390
Percentile: 49.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: webdev-react. Source rank: #39. Votes: 6716. Organization: openai. License: Proprietary.

49.2% percentile inside its fair comparison set

1,390Raw benchmark valueCI 1,382 - 1,398

Code Migration

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #17 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 12.9%
Percentile: 27.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: code-migration; provider: OpenAI.

27.3% percentile inside its fair comparison set

12.9%Raw benchmark valueCI 5.8% - 20.1%

LiveCodeBench

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #10 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 86.6%
Percentile: 90%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: lcb; provider: OpenAI.

90% percentile inside its fair comparison set

86.6%Raw benchmark valueCI 84.7% - 88.5%

ProgramBench

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #8 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 0%
Percentile: 70%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: programbench; provider: OpenAI.

70% percentile inside its fair comparison set

0%Raw benchmark valueCI 0% - 0%

SWE-bench Verified

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #27 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 73%
Percentile: 51.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: swebench; provider: OpenAI.

51.9% percentile inside its fair comparison set

73%Raw benchmark valueCI 69.1% - 76.9%

Terminal-Bench 2.1

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #14 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 54.7%
Percentile: 51.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: terminal-bench-2-1; provider: OpenAI.

51.9% percentile inside its fair comparison set

54.7%Raw benchmark valueCI 45.8% - 63.6%

Vibe Code Bench v1.1

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #13 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 48%
Percentile: 75.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: vibe-code; provider: OpenAI.

75.5% percentile inside its fair comparison set

48%Raw benchmark valueCI 37% - 59%

Text Arena · Coding

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #39 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,496
Percentile: 88.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: coding. Source rank: #49. Votes: 10646. Organization: openai. License: Proprietary.

88.1% percentile inside its fair comparison set

1,496Raw benchmark valueCI 1,489 - 1,502

Text Arena · Coding · No Style Control

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #65 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,438
Percentile: 80%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: coding. Source rank: #80. Votes: 10646. Organization: openai. License: Proprietary.

80% percentile inside its fair comparison set

1,438Raw benchmark valueCI 1,431 - 1,445

IOI

VALS-AI · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #28 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 6.8%
Percentile: 38.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: ioi; provider: OpenAI.

38.6% percentile inside its fair comparison set

6.8%Raw benchmark valueCI 0.3% - 13.2%

Agentic coding

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #92 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 17%
Percentile: 15.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Agentic Coding. Tasks scored: 3.

15.7% percentile inside its fair comparison set

17%Raw benchmark value

Coding

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #82 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 68.2%
Percentile: 25%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Coding. Tasks scored: 2.

25% percentile inside its fair comparison set

68.2%Raw benchmark value

JavaScript

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #84 · Source label: gpt-5.4-mini-medium

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 21.1%
Percentile: 23.1%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: javascript. Category: Agentic Coding.

23.1% percentile inside its fair comparison set

21.1%Raw benchmark value

TypeScript

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #91 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 10%
Percentile: 16.8%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: typescript. Category: Agentic Coding.

16.8% percentile inside its fair comparison set

10%Raw benchmark value

Python

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #98 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 15%
Percentile: 10.2%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: python. Category: Agentic Coding.

10.2% percentile inside its fair comparison set

15%Raw benchmark value

Coding generation

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #89 · Source label: gpt-5.4-mini-xhigh

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 67.7%
Percentile: 18.5%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: code_generation. Category: Coding.

18.5% percentile inside its fair comparison set

67.7%Raw benchmark value

Coding completion

LB · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #88 · Source label: gpt-5.4-mini-medium

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 64.4%
Percentile: 19.4%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: code_completion. Category: Coding.

19.4% percentile inside its fair comparison set

64.4%Raw benchmark value

Terminal-Bench 2.0

TERMINAL-BENCH · Coding · Objective

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #20 · Source label: gpt-5-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Terminal-Bench
Raw value: 31.9%
Percentile: 36.7%
Last updated: archived
Eligibility: headline eligible

36.7% percentile inside its fair comparison set

31.9%Raw benchmark value

Reasoning / math / science21 benchmarks35.8%

Humanity's Last Exam

AA · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #219 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 5%
Percentile: 41.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `hle`.

41.1% percentile inside its fair comparison set

5%Raw benchmark value

GPQA

AA · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #197 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 60.6%
Percentile: 47.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `gpqa`.

47.6% percentile inside its fair comparison set

60.6%Raw benchmark value

CritPt

AA · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #106 · Source label: GPT-5 mini (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 0%
Percentile: 65.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `critpt`.

65.2% percentile inside its fair comparison set

0%Raw benchmark value

ProofBench

VALS-AI · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #15 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 21%
Percentile: 60%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: proof_bench; provider: OpenAI.

60% percentile inside its fair comparison set

21%Raw benchmark valueCI 13% - 29%

GPQA Diamond

VALS-AI · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #36 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 83.1%
Percentile: 60.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: gpqa; provider: OpenAI.

60.7% percentile inside its fair comparison set

83.1%Raw benchmark valueCI 78.3% - 87.9%

MMLU Pro

VALS-AI · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #38 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 84.6%
Percentile: 58.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: mmlu_pro; provider: OpenAI.

58.4% percentile inside its fair comparison set

84.6%Raw benchmark valueCI 83.9% - 85.3%

Text Arena · Math

AR · Reasoning / math / science · Human

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #43 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,441
Percentile: 86.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: math. Source rank: #54. Votes: 2233. Organization: openai. License: Proprietary.

86.6% percentile inside its fair comparison set

1,441Raw benchmark valueCI 1,428 - 1,454

Text Arena · Math · No Style Control

AR · Reasoning / math / science · Human

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #69 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,421
Percentile: 78.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: math. Source rank: #83. Votes: 2233. Organization: openai. License: Proprietary.

78.3% percentile inside its fair comparison set

1,421Raw benchmark valueCI 1,408 - 1,434

TutorBench

SL · Reasoning / math / science · Rubric

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 55.3%
Percentile: 80%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Manually verified from the official Scale Labs TutorBench leaderboard. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

80% percentile inside its fair comparison set

55.3%Raw benchmark value

MultiNRC

SL · Reasoning / math / science · Rubric

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #6 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 52.1%
Percentile: 60%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

60% percentile inside its fair comparison set

52.1%Raw benchmark value

EnigmaEval

SL · Reasoning / math / science · Rubric

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #6 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 64%
Percentile: 80%
Last updated: aging
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

80% percentile inside its fair comparison set

64%Raw benchmark value

Mathematics

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #107 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 37%
Percentile: 1.9%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Mathematics. Tasks scored: 4.

1.9% percentile inside its fair comparison set

37%Raw benchmark value

Reasoning

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #107 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 21.9%
Percentile: 1.9%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Reasoning. Tasks scored: 4.

1.9% percentile inside its fair comparison set

21.9%Raw benchmark value

AMPS Hard

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #104 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 68%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: AMPS_Hard. Category: Mathematics.

4.6% percentile inside its fair comparison set

68%Raw benchmark value

Integrals with game

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #104 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 6%
Percentile: 4.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: integrals_with_game. Category: Mathematics.

4.6% percentile inside its fair comparison set

6%Raw benchmark value

Math competition

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #109 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 18.6%
Percentile: 0%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: math_comp. Category: Mathematics.

0% percentile inside its fair comparison set

18.6%Raw benchmark value

Olympiad

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #103 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 51.3%
Percentile: 5.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: olympiad. Category: Mathematics.

5.6% percentile inside its fair comparison set

51.3%Raw benchmark value

Theory of mind

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #105 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 28.8%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: theory_of_mind. Category: Reasoning.

3.7% percentile inside its fair comparison set

28.8%Raw benchmark value

Zebra puzzle

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #103 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 6.7%
Percentile: 4.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: zebra_puzzle. Category: Reasoning.

4.7% percentile inside its fair comparison set

6.7%Raw benchmark value

Spatial

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #105 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 46%
Percentile: 3.7%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: spatial. Category: Reasoning.

3.7% percentile inside its fair comparison set

46%Raw benchmark value

Logic with navigation

LB · Reasoning / math / science · Objective

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #106 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 6%
Percentile: 2.8%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: logic_with_navigation. Category: Reasoning.

2.8% percentile inside its fair comparison set

6%Raw benchmark value

Professional reasoning35 benchmarks64.4%

GDPval-AA

AA · Professional reasoning · Rubric

Agentic performance on economically valuable work tasks.

Rank #35 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 755
Percentile: 26.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `gdpvalBreakdown.elo`.

26.1% percentile inside its fair comparison set

755Raw benchmark value

APEX-Agents-AA

AA · Professional reasoning · Objective

Long-horizon agentic task completion.

Rank #7 · Source label: GPT-5.4 mini (xhigh)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 28.2%
Percentile: 75%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `apexAgents`.

75% percentile inside its fair comparison set

28.2%Raw benchmark value

Legal Research Bench

VALS-AI · Professional reasoning · Objective

Applied legal research tasks.

Rank #13 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 12.5%
Percentile: 0%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: legal_research; provider: OpenAI.

0% percentile inside its fair comparison set

12.5%Raw benchmark valueCI 8% - 17%

Vals Index

VALS-AI · Professional reasoning · Combined

Weighted model performance across economically relevant Vals tasks.

Rank #14 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 52
Percentile: 50%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: vals_index; provider: OpenAI.

50% percentile inside its fair comparison set

52Raw benchmark valueCI 48 - 56

LegalBench

VALS-AI · Professional reasoning · Objective

Academic legal reasoning tasks.

Rank #47 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 81.8%
Percentile: 48.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: legal_bench; provider: OpenAI.

48.9% percentile inside its fair comparison set

81.8%Raw benchmark valueCI 80.9% - 82.7%

Finance Agent v2

VALS-AI · Professional reasoning · Objective

Core financial analyst tasks for agentic models.

Rank #9 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 45.4%
Percentile: 68%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: fabv2; provider: OpenAI.

68% percentile inside its fair comparison set

45.4%Raw benchmark valueCI 44.5% - 46.2%

TaxEval v2

VALS-AI · Professional reasoning · Objective

Answer quality on tax questions and responses.

Rank #11 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 75.2%
Percentile: 89%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: tax_eval_v2; provider: OpenAI.

89% percentile inside its fair comparison set

75.2%Raw benchmark valueCI 73.6% - 76.9%

MedCode

VALS-AI · Professional reasoning · Objective

Medical billing support and coding tasks.

Rank #22 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 43%
Percentile: 58.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: medcode; provider: OpenAI.

58.8% percentile inside its fair comparison set

43%Raw benchmark valueCI 39% - 47.1%

MedScribe

VALS-AI · Professional reasoning · Objective

Administrative documentation support for doctors.

Rank #19 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 80.6%
Percentile: 64%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: medscribe; provider: OpenAI.

64% percentile inside its fair comparison set

80.6%Raw benchmark valueCI 76.8% - 84.3%

Text Arena · Expert

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena expert leaderboard.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 88.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: expert. Source rank: #39. Votes: 3590. Organization: openai. License: Proprietary.

88.7% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,471 - 1,492

Text Arena · Industry Business And Management And Financial Operations

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_business_and_management_and_financial_operations leaderboard.

Rank #22 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,460
Percentile: 93.4%
Last updated: recent
Eligibility: headline eligible

93.4% percentile inside its fair comparison set

1,460Raw benchmark valueCI 1,452 - 1,468

Text Arena · Industry Entertainment And Sports And Media

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_entertainment_and_sports_and_media leaderboard.

Rank #47 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,408
Percentile: 85.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_entertainment_and_sports_and_media. Source rank: #59. Votes: 8212. Organization: openai. License: Proprietary.

85.8% percentile inside its fair comparison set

1,408Raw benchmark valueCI 1,400 - 1,415

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_legal_and_government leaderboard.

Rank #41 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,457
Percentile: 86.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_legal_and_government. Source rank: #54. Votes: 3015. Organization: openai. License: Proprietary.

86.6% percentile inside its fair comparison set

1,457Raw benchmark valueCI 1,445 - 1,469

Text Arena · Industry Life And Physical And Social Science

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_life_and_physical_and_social_science leaderboard.

Rank #41 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,465
Percentile: 87.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_life_and_physical_and_social_science. Source rank: #52. Votes: 6537. Organization: openai. License: Proprietary.

87.6% percentile inside its fair comparison set

1,465Raw benchmark valueCI 1,457 - 1,474

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_mathematical leaderboard.

Rank #37 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,456
Percentile: 88.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_mathematical. Source rank: #45. Votes: 2140. Organization: openai. License: Proprietary.

88.3% percentile inside its fair comparison set

1,456Raw benchmark valueCI 1,443 - 1,470

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_medicine_and_healthcare leaderboard.

Rank #55 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,455
Percentile: 81.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_medicine_and_healthcare. Source rank: #71. Votes: 2894. Organization: openai. License: Proprietary.

81.7% percentile inside its fair comparison set

1,455Raw benchmark valueCI 1,443 - 1,467

Text Arena · Industry Software And It Services

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_software_and_it_services leaderboard.

Rank #38 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 88.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_software_and_it_services. Source rank: #49. Votes: 15521. Organization: openai. License: Proprietary.

88.6% percentile inside its fair comparison set

1,487Raw benchmark valueCI 1,481 - 1,493

Text Arena · Industry Writing And Literature And Language

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_writing_and_literature_and_language leaderboard.

Rank #44 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,423
Percentile: 86.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_writing_and_literature_and_language. Source rank: #57. Votes: 9541. Organization: openai. License: Proprietary.

86.7% percentile inside its fair comparison set

1,423Raw benchmark valueCI 1,416 - 1,430

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena expert leaderboard.

Rank #57 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,436
Percentile: 79.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: expert. Source rank: #68. Votes: 3590. Organization: openai. License: Proprietary.

79.6% percentile inside its fair comparison set

1,436Raw benchmark valueCI 1,425 - 1,446

Text Arena · Industry Business And Management And Financial Operations · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_business_and_management_and_financial_operations leaderboard.

Rank #59 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,417
Percentile: 81.8%
Last updated: recent
Eligibility: headline eligible

81.8% percentile inside its fair comparison set

1,417Raw benchmark valueCI 1,410 - 1,425

Text Arena · Industry Entertainment And Sports And Media · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_entertainment_and_sports_and_media leaderboard.

Rank #82 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,372
Percentile: 74.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_entertainment_and_sports_and_media. Source rank: #100. Votes: 8212. Organization: openai. License: Proprietary.

74.9% percentile inside its fair comparison set

1,372Raw benchmark valueCI 1,364 - 1,379

Text Arena · Industry Legal And Government · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_legal_and_government leaderboard.

Rank #83 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,415
Percentile: 72.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_legal_and_government. Source rank: #101. Votes: 3015. Organization: openai. License: Proprietary.

72.5% percentile inside its fair comparison set

1,415Raw benchmark valueCI 1,403 - 1,427

Text Arena · Industry Life And Physical And Social Science · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_life_and_physical_and_social_science leaderboard.

Rank #91 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,417
Percentile: 72.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_life_and_physical_and_social_science. Source rank: #107. Votes: 6537. Organization: openai. License: Proprietary.

72.1% percentile inside its fair comparison set

1,417Raw benchmark valueCI 1,409 - 1,425

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_mathematical leaderboard.

Rank #58 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 81.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_mathematical. Source rank: #68. Votes: 2140. Organization: openai. License: Proprietary.

81.5% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,419 - 1,446

Text Arena · Industry Medicine And Healthcare · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_medicine_and_healthcare leaderboard.

Rank #101 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,397
Percentile: 66.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_medicine_and_healthcare. Source rank: #122. Votes: 2894. Organization: openai. License: Proprietary.

66.1% percentile inside its fair comparison set

1,397Raw benchmark valueCI 1,385 - 1,409

Text Arena · Industry Software And It Services · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_software_and_it_services leaderboard.

Rank #71 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,436
Percentile: 78.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_software_and_it_services. Source rank: #86. Votes: 15521. Organization: openai. License: Proprietary.

78.5% percentile inside its fair comparison set

1,436Raw benchmark valueCI 1,430 - 1,442

Text Arena · Industry Writing And Literature And Language · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_writing_and_literature_and_language leaderboard.

Rank #71 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,392
Percentile: 78.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: industry_writing_and_literature_and_language. Source rank: #86. Votes: 9541. Organization: openai. License: Proprietary.

78.4% percentile inside its fair comparison set

1,392Raw benchmark valueCI 1,385 - 1,399

SAGE

VALS-AI · Professional reasoning · Objective

Student Assessment with Generative Evaluation.

Rank #9 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 50.8%
Percentile: 82.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: sage; provider: OpenAI.

82.2% percentile inside its fair comparison set

50.8%Raw benchmark valueCI 44.1% - 57.5%

PRBench Legal

SL · Professional reasoning · Rubric

Applied legal reasoning on professional-domain tasks.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 49%
Percentile: 83.3%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

83.3% percentile inside its fair comparison set

49%Raw benchmark value

Data analysis

LB · Professional reasoning · Objective

Structured data manipulation and table reasoning accuracy.

Rank #101 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 41.5%
Percentile: 7.4%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category: Data Analysis. Tasks scored: 3.

7.4% percentile inside its fair comparison set

41.5%Raw benchmark value

Overall

LB · Professional reasoning · Objective

Average objective performance across LiveBench's current public category mix.

Rank #103 · Source label: gpt-5.4-mini

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 37%
Percentile: 5.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Category averages included: 7.

5.6% percentile inside its fair comparison set

37%Raw benchmark value

Consecutive events

LB · Professional reasoning · Objective

Objective consecutive events score in LiveBench.

Rank #103 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 0.5%
Percentile: 5.6%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: consecutive_events. Category: Data Analysis.

5.6% percentile inside its fair comparison set

0.5%Raw benchmark value

Table join

LB · Professional reasoning · Objective

Objective table join score in LiveBench.

Rank #99 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 29.9%
Percentile: 9.3%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: tablejoin. Category: Data Analysis.

9.3% percentile inside its fair comparison set

29.9%Raw benchmark value

Table reformat

LB · Professional reasoning · Objective

Objective table reformat score in LiveBench.

Rank #93 · Source label: gpt-5-mini-minimal

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 94.1%
Percentile: 14.8%
Last updated: archived
Eligibility: headline eligible

Derived from the official LiveBench website leaderboard table. Task: tablereformat. Category: Data Analysis.

14.8% percentile inside its fair comparison set

94.1%Raw benchmark value

Poker Agent

VALS-AI · Professional reasoning · Objective

Agent profit in poker-style strategic play.

Rank #4 · Source label: openai/gpt-5-2025-08-07

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 1,103.2 score
Percentile: 93.8%
Last updated: archived
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Parsed from Vals AI BenchmarkView overall scores. Vals slug: poker_agent; provider: unknown. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

93.8% percentile inside its fair comparison set

1,103.2 scoreRaw benchmark valueCI 1,103.2 score - 1,103.2 score

Search / tool use3 benchmarks31.7%

Tau2-Bench Telecom

AA · Search / tool use · Objective

It matters when the model must browse, call tools, and recover useful answers from external systems.

Rank #212 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.4%
Percentile: 31.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `tau2`.

31.7% percentile inside its fair comparison set

23.4%Raw benchmark value

Search Arena

AR · Search / tool use · Human

It matters when the model must browse, call tools, and recover useful answers from external systems.

Rank #19 · Source label: gpt-5-search

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,180
Percentile: 43.3%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

43.3% percentile inside its fair comparison set

1,180Raw benchmark valueCI 1,173 - 1,188

Search Arena · No Style Control

AR · Search / tool use · Human

It matters when the model must browse, call tools, and recover useful answers from external systems.

Rank #26 · Source label: gpt-5-search

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,132
Percentile: 20%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

20% percentile inside its fair comparison set

1,132Raw benchmark valueCI 1,127 - 1,138

Long context2 benchmarks50%

Long Context Reasoning

AA · Long context · Objective

It checks whether long-context claims survive contact with retrieval, memory, or long-document tasks.

Rank #155 · Source label: GPT-5.4 mini (Non-Reasoning)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 30.3%
Percentile: 51.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `lcr`.

51.1% percentile inside its fair comparison set

30.3%Raw benchmark value

CorpFin v2

VALS-AI · Long context · Objective

It checks whether long-context claims survive contact with retrieval, memory, or long-document tasks.

Rank #46 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 60.9%
Percentile: 48.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: corp_fin_v2; provider: OpenAI.

48.9% percentile inside its fair comparison set

60.9%Raw benchmark valueCI 59% - 62.8%

Vision understanding24 benchmarks56.9%

MMMU-Pro

AA · Vision understanding · Objective

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #81 · Source label: GPT-5 mini (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 58.4%
Percentile: 40.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `mmmuPro`.

40.7% percentile inside its fair comparison set

58.4%Raw benchmark value

Vision Arena

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #17 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,256
Percentile: 85.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #23. Votes: 15692. Organization: openai. License: Proprietary.

85.3% percentile inside its fair comparison set

1,256Raw benchmark valueCI 1,249 - 1,263

Vision Arena · Captioning

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,165
Percentile: 26.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: captioning. Source rank: #21. Votes: 302. Organization: openai. License: Proprietary.

26.9% percentile inside its fair comparison set

1,165Raw benchmark valueCI 1,125 - 1,206

Vision Arena · Creative Writing Vision

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #29 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,233
Percentile: 49.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing_vision. Source rank: #36. Votes: 897. Organization: openai. License: Proprietary.

49.1% percentile inside its fair comparison set

1,233Raw benchmark valueCI 1,212 - 1,254

Vision Arena · Diagram

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #17 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,281
Percentile: 77.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: diagram. Source rank: #23. Votes: 4259. Organization: openai. License: Proprietary.

77.1% percentile inside its fair comparison set

1,281Raw benchmark valueCI 1,270 - 1,292

Vision Arena · English

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,252
Percentile: 82.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #25. Votes: 6782. Organization: openai. License: Proprietary.

82.6% percentile inside its fair comparison set

1,252Raw benchmark valueCI 1,242 - 1,262

Vision Arena · Entity Recognition

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,205
Percentile: 40.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: entity_recognition. Source rank: #19. Votes: 283. Organization: openai. License: Proprietary.

40.6% percentile inside its fair comparison set

1,205Raw benchmark valueCI 1,165 - 1,245

Vision Arena · Homework

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #16 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,290
Percentile: 77.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: homework. Source rank: #21. Votes: 2433. Organization: openai. License: Proprietary.

77.9% percentile inside its fair comparison set

1,290Raw benchmark valueCI 1,277 - 1,303

Vision Arena · Humor

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #22 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,237
Percentile: 57.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: humor. Source rank: #28. Votes: 498. Organization: openai. License: Proprietary.

57.1% percentile inside its fair comparison set

1,237Raw benchmark valueCI 1,209 - 1,265

Vision Arena · Ocr

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #18 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,267
Percentile: 75.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: ocr. Source rank: #24. Votes: 11142. Organization: openai. License: Proprietary.

75.7% percentile inside its fair comparison set

1,267Raw benchmark valueCI 1,259 - 1,274

Vision Arena · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,250
Percentile: 75.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: overall. Source rank: #35. Votes: 15692. Organization: openai. License: Proprietary.

75.2% percentile inside its fair comparison set

1,250Raw benchmark valueCI 1,242 - 1,257

Vision Arena · Captioning · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #16 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,213
Percentile: 42.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: captioning. Source rank: #15. Votes: 302. Organization: openai. License: Proprietary.

42.3% percentile inside its fair comparison set

1,213Raw benchmark valueCI 1,179 - 1,247

Vision Arena · Creative Writing Vision · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #36 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,227
Percentile: 36.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: creative_writing_vision. Source rank: #44. Votes: 897. Organization: openai. License: Proprietary.

36.4% percentile inside its fair comparison set

1,227Raw benchmark valueCI 1,206 - 1,248

Vision Arena · Diagram · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,263
Percentile: 61.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: diagram. Source rank: #35. Votes: 4259. Organization: openai. License: Proprietary.

61.4% percentile inside its fair comparison set

1,263Raw benchmark valueCI 1,252 - 1,273

Vision Arena · English · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #32 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,246
Percentile: 71.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: english. Source rank: #39. Votes: 6782. Organization: openai. License: Proprietary.

71.6% percentile inside its fair comparison set

1,246Raw benchmark valueCI 1,236 - 1,256

Vision Arena · Entity Recognition · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #20 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,225
Percentile: 40.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: entity_recognition. Source rank: #19. Votes: 283. Organization: openai. License: Proprietary.

40.6% percentile inside its fair comparison set

1,225Raw benchmark valueCI 1,191 - 1,259

Vision Arena · Homework · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #21 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,285
Percentile: 70.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: homework. Source rank: #27. Votes: 2433. Organization: openai. License: Proprietary.

70.6% percentile inside its fair comparison set

1,285Raw benchmark valueCI 1,272 - 1,298

Vision Arena · Humor · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #27 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,239
Percentile: 46.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: humor. Source rank: #33. Votes: 498. Organization: openai. License: Proprietary.

46.9% percentile inside its fair comparison set

1,239Raw benchmark valueCI 1,211 - 1,267

Vision Arena · Ocr · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,256
Percentile: 61.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: ocr. Source rank: #36. Votes: 11142. Organization: openai. License: Proprietary.

61.4% percentile inside its fair comparison set

1,256Raw benchmark valueCI 1,249 - 1,264

MMMU Pro

VALS-AI · Vision understanding · Objective

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #31 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 79.2%
Percentile: 48.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: mmmu; provider: OpenAI.

48.3% percentile inside its fair comparison set

79.2%Raw benchmark valueCI 77.3% - 81.2%

VTB

SL · Vision understanding · Rubric

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #6 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 17%
Percentile: 63.6%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

63.6% percentile inside its fair comparison set

17%Raw benchmark value

VISTA

SL · Vision understanding · Rubric

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 79%
Percentile: 92.9%
Last updated: aging
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

92.9% percentile inside its fair comparison set

79%Raw benchmark value

Vision Arena · Creative Writing

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #28 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,152
Percentile: 15.6%
Last updated: archived
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: creative_writing. Source rank: #29. Votes: 1108. Organization: openai. License: Proprietary.

15.6% percentile inside its fair comparison set

1,152Raw benchmark valueCI 1,132 - 1,173

Vision Arena · Creative Writing · No Style Control

AR · Vision understanding · Human

It is useful when the model must read charts, UI, screenshots, or visual scenes rather than text alone.

Rank #25 · Source label: gpt-5-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,181
Percentile: 25%
Last updated: archived
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5-mini-high`. Category: creative_writing. Source rank: #26. Votes: 1108. Organization: openai. License: Proprietary.

25% percentile inside its fair comparison set

1,181Raw benchmark valueCI 1,163 - 1,199

Document understanding3 benchmarks70.7%

Vals Multimodal Index

VALS-AI · Document understanding · Combined

It matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.

Rank #10 · Source label: openai/gpt-5.4-mini-2026-03-17

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 54
Percentile: 52.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: vals_multimodal_index; provider: OpenAI.

52.6% percentile inside its fair comparison set

54Raw benchmark valueCI 51 - 58

MortgageTax

VALS-AI · Document understanding · Objective

It matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.

Rank #21 · Source label: openai/gpt-5-mini-2025-08-07

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 66.9%
Percentile: 66.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: mortgage_tax; provider: OpenAI.

66.7% percentile inside its fair comparison set

66.9%Raw benchmark valueCI 65.1% - 68.7%

Multimodal mix

OC · Document understanding · Objective

It matters when the job is reading PDFs, tables, forms, or mixed-layout documents rather than plain chat.

Rank #4 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: OpenCompass
Raw value: 75.4%
Percentile: 92.9%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

92.9% percentile inside its fair comparison set

75.4%Raw benchmark value

Safety1 benchmark61.5%

MASK

SL · Safety · Rubric

Whether a model stays honest instead of covertly optimizing against the user.

Rank #7 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 79.3%
Percentile: 61.5%
Last updated: recent
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

61.5% percentile inside its fair comparison set

79.3%Raw benchmark value

Embeddings / retrieval1 benchmark100%

Retrieval

MTEB · Embeddings / retrieval · Retrieval

It is one of the few direct signals for retrieval stacks, where embedding quality matters more than chat style.

Rank #3 · Source label: gpt-5

backfilledproxy backfilledBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: MTEB
Raw value: 58.8 ndcg
Percentile: 100%
Last updated: aging
Eligibility: Fallback benchmark identity is visible for context but excluded from default ranking.

Embedding endpoint score. Backfilled from GPT-5 via approved benchmark identity mapping map-gpt-5-4-mini-to-gpt-5.

100% percentile inside its fair comparison set

58.8 ndcgRaw benchmark value

Multilingual16 benchmarks77.4%

Text Arena · Chinese

AR · Multilingual · Human

Observed user preference in Arena's Text Arena chinese leaderboard.

Rank #50 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 83.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #61. Votes: 2151. Organization: openai. License: Proprietary.

83.4% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,465 - 1,493

Text Arena · French

AR · Multilingual · Human

Observed user preference in Arena's Text Arena french leaderboard.

Rank #41 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,465
Percentile: 81.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: french. Source rank: #52. Votes: 1303. Organization: openai. License: Proprietary.

81.5% percentile inside its fair comparison set

1,465Raw benchmark valueCI 1,446 - 1,484

Text Arena · German

AR · Multilingual · Human

Observed user preference in Arena's Text Arena german leaderboard.

Rank #42 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,437
Percentile: 82.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: german. Source rank: #57. Votes: 537. Organization: openai. License: Proprietary.

82.7% percentile inside its fair comparison set

1,437Raw benchmark valueCI 1,411 - 1,463

Text Arena · Japanese

AR · Multilingual · Human

Observed user preference in Arena's Text Arena japanese leaderboard.

Rank #31 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,415
Percentile: 85.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: japanese. Source rank: #43. Votes: 281. Organization: openai. License: Proprietary.

85.2% percentile inside its fair comparison set

1,415Raw benchmark valueCI 1,377 - 1,452

Text Arena · Korean

AR · Multilingual · Human

Observed user preference in Arena's Text Arena korean leaderboard.

Rank #51 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,384
Percentile: 76%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: korean. Source rank: #65. Votes: 607. Organization: openai. License: Proprietary.

76% percentile inside its fair comparison set

1,384Raw benchmark valueCI 1,358 - 1,409

Text Arena · Russian

AR · Multilingual · Human

Observed user preference in Arena's Text Arena russian leaderboard.

Rank #34 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,453
Percentile: 88.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: russian. Source rank: #43. Votes: 4221. Organization: openai. License: Proprietary.

88.6% percentile inside its fair comparison set

1,453Raw benchmark valueCI 1,443 - 1,462

Text Arena · Spanish

AR · Multilingual · Human

Observed user preference in Arena's Text Arena spanish leaderboard.

Rank #39 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,443
Percentile: 82.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: spanish. Source rank: #49. Votes: 1135. Organization: openai. License: Proprietary.

82.2% percentile inside its fair comparison set

1,443Raw benchmark valueCI 1,423 - 1,462

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena chinese leaderboard.

Rank #77 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,451
Percentile: 74.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #93. Votes: 2151. Organization: openai. License: Proprietary.

74.2% percentile inside its fair comparison set

1,451Raw benchmark valueCI 1,437 - 1,465

Text Arena · French · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena french leaderboard.

Rank #63 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,431
Percentile: 71.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: french. Source rank: #78. Votes: 1303. Organization: openai. License: Proprietary.

71.3% percentile inside its fair comparison set

1,431Raw benchmark valueCI 1,412 - 1,449

Text Arena · German · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena german leaderboard.

Rank #70 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,403
Percentile: 70.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: german. Source rank: #87. Votes: 537. Organization: openai. License: Proprietary.

70.9% percentile inside its fair comparison set

1,403Raw benchmark valueCI 1,377 - 1,428

Text Arena · Japanese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena japanese leaderboard.

Rank #46 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,383
Percentile: 77.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: japanese. Source rank: #59. Votes: 281. Organization: openai. License: Proprietary.

77.8% percentile inside its fair comparison set

1,383Raw benchmark valueCI 1,345 - 1,421

Text Arena · Korean · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena korean leaderboard.

Rank #74 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,358
Percentile: 64.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: korean. Source rank: #90. Votes: 607. Organization: openai. License: Proprietary.

64.9% percentile inside its fair comparison set

1,358Raw benchmark valueCI 1,333 - 1,384

Text Arena · Russian · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena russian leaderboard.

Rank #55 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,419
Percentile: 81.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: russian. Source rank: #69. Votes: 4221. Organization: openai. License: Proprietary.

81.3% percentile inside its fair comparison set

1,419Raw benchmark valueCI 1,409 - 1,429

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena spanish leaderboard.

Rank #66 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,413
Percentile: 69.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: spanish. Source rank: #80. Votes: 1135. Organization: openai. License: Proprietary.

69.6% percentile inside its fair comparison set

1,413Raw benchmark valueCI 1,393 - 1,432

Vision Arena · Chinese

AR · Multilingual · Human

Observed user preference in Arena's Vision Arena chinese leaderboard.

Rank #17 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,301
Percentile: 79.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #22. Votes: 899. Organization: openai. License: Proprietary.

79.2% percentile inside its fair comparison set

1,301Raw benchmark valueCI 1,276 - 1,326

Vision Arena · Chinese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Vision Arena chinese leaderboard.

Rank #24 · Source label: gpt-5.4-mini-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,302
Percentile: 70.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gpt-5.4-mini-high`. Category: chinese. Source rank: #29. Votes: 899. Organization: openai. License: Proprietary.

70.1% percentile inside its fair comparison set

1,302Raw benchmark valueCI 1,278 - 1,327