Model profile · Google

Gemini 3.5 Flash

Closed weightsfrontier · registry tag 2026 current

Visible tradeoffs

Reads as visible tradeoffs across the resolved source data.

Visible coverage: 55.5%
Verified coverage: 55.5%
Spread: 92.4%
Last verified: Jun 20, 2026

78%bench fit

textcodevisiondocument9 aliases31 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 / text37 benchmarks84.4%

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 #30 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 35
Percentile: 92.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `intelligenceIndex`.

92.7% percentile inside its fair comparison set

35Raw 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 #12 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 43.1%
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceAccuracy`.

96.3% percentile inside its fair comparison set

43.1%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 #75 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 26.6%
Percentile: 75.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceNonHallucination`.

75.2% percentile inside its fair comparison set

26.6%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 #105 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 47.3%
Percentile: 67%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `ifbench`.

67% percentile inside its fair comparison set

47.3%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 #226 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $3.4 /1M tokens
Percentile: 18.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mBlended0To3To1`.

18.5% percentile inside its fair comparison set

$3.4 /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 #220 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $1.5 /1M input tokens
Percentile: 20.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mInputTokens`.

20.7% percentile inside its fair comparison set

$1.5 /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 #234 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $9 /1M output tokens
Percentile: 15.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mOutputTokens`.

15.6% percentile inside its fair comparison set

$9 /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 #36 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 177.8 tokens/s
Percentile: 83.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianOutputTokensPerSecond`.

83.3% percentile inside its fair comparison set

177.8 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 #195 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.61s
Percentile: 7.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstTokenSeconds`.

7.6% percentile inside its fair comparison set

23.61sRaw 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 #142 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.61s
Percentile: 32.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstAnswerTokenSeconds`.

32.9% percentile inside its fair comparison set

23.61sRaw 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 #11 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,476
Percentile: 96.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: overall. Source rank: #13. Votes: 10171. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,476Raw benchmark valueCI 1,470 - 1,483

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 #8 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,464
Percentile: 97.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: creative_writing. Source rank: #10. Votes: 1524. Organization: google. License: Proprietary.

97.8% percentile inside its fair comparison set

1,464Raw benchmark valueCI 1,448 - 1,480

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 #19 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,477
Percentile: 94.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: english. Source rank: #24. Votes: 4948. Organization: google. License: Proprietary.

94.5% percentile inside its fair comparison set

1,477Raw benchmark valueCI 1,468 - 1,486

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,484
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: exclude_ties. Source rank: #15. Votes: 7856. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,484Raw benchmark valueCI 1,476 - 1,493

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 #18 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 94.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts. Source rank: #23. Votes: 6841. Organization: google. License: Proprietary.

94.8% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,484 - 1,499

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 #28 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,484
Percentile: 91.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts_english. Source rank: #36. Votes: 3483. Organization: google. License: Proprietary.

91.7% percentile inside its fair comparison set

1,484Raw benchmark valueCI 1,473 - 1,495

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 #18 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 94.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: instruction_following. Source rank: #25. Votes: 3458. Organization: google. License: Proprietary.

94.8% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,451 - 1,473

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 #18 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,480
Percentile: 94.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: longer_query. Source rank: #24. Votes: 4434. Organization: google. License: Proprietary.

94.4% percentile inside its fair comparison set

1,480Raw benchmark valueCI 1,470 - 1,490

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 #11 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 96.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: multi_turn. Source rank: #13. Votes: 1848. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,487Raw benchmark valueCI 1,473 - 1,502

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 #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,480
Percentile: 98.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: overall. Source rank: #7. Votes: 10171. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,480Raw benchmark valueCI 1,473 - 1,486

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 #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 98.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: creative_writing. Source rank: #8. Votes: 1524. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,453 - 1,485

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 #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 98.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: english. Source rank: #9. Votes: 4948. Organization: google. License: Proprietary.

98.2% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,470 - 1,488

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 #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,486
Percentile: 98.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: exclude_ties. Source rank: #8. Votes: 7856. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,486Raw benchmark valueCI 1,478 - 1,495

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 #9 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,484
Percentile: 97.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts. Source rank: #11. Votes: 6841. Organization: google. License: Proprietary.

97.5% percentile inside its fair comparison set

1,484Raw benchmark valueCI 1,476 - 1,492

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 #17 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 95.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts_english. Source rank: #20. Votes: 3483. Organization: google. License: Proprietary.

95.1% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,467 - 1,489

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,463
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: instruction_following. Source rank: #17. Votes: 3458. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,463Raw benchmark valueCI 1,452 - 1,473

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 #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,481
Percentile: 98%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: longer_query. Source rank: #10. Votes: 4434. Organization: google. License: Proprietary.

98% percentile inside its fair comparison set

1,481Raw benchmark valueCI 1,472 - 1,491

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 #4 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,488
Percentile: 99.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: multi_turn. Source rank: #6. Votes: 1848. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,488Raw benchmark valueCI 1,473 - 1,502

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 #2 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75.6%
Percentile: 99.1%
Last updated: archived
Eligibility: headline eligible

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

99.1% percentile inside its fair comparison set

75.6%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 #6 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 84.6%
Percentile: 95.4%
Last updated: archived
Eligibility: headline eligible

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

95.4% percentile inside its fair comparison set

84.6%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 #3 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75.5%
Percentile: 98.1%
Last updated: archived
Eligibility: headline eligible

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

98.1% percentile inside its fair comparison set

75.5%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 #3 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 69.1%
Percentile: 98.1%
Last updated: archived
Eligibility: headline eligible

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

98.1% percentile inside its fair comparison set

69.1%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 #4 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75.2%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

75.2%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 #4 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 82.6%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

82.6%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 #5 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 100%
Percentile: 100%
Last updated: archived
Eligibility: headline eligible

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

100% percentile inside its fair comparison set

100%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 #7 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 67.8%
Percentile: 94.4%
Last updated: archived
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

67.8%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 #5 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 86%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

86%Raw benchmark value

Coding23 benchmarks83.2%

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 #23 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 39.4%
Percentile: 92.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `terminalbenchHard`.

92.7% percentile inside its fair comparison set

39.4%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 #11 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 48.8%
Percentile: 97.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `scicode`.

97.3% percentile inside its fair comparison set

48.8%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 #5 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 70
Percentile: 94.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `codingIndex`.

94.7% percentile inside its fair comparison set

70Raw 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 #7 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 37
Percentile: 87%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `agenticIndex`.

87% percentile inside its fair comparison set

37Raw 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 #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,506
Percentile: 87.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: overall. Source rank: #14. Votes: 2217. Organization: google. License: Proprietary.

87.7% percentile inside its fair comparison set

1,506Raw benchmark valueCI 1,492 - 1,519

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 #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,506
Percentile: 87.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: webdev. Source rank: #14. Votes: 2217. Organization: google. License: Proprietary.

87.7% percentile inside its fair comparison set

1,506Raw benchmark valueCI 1,492 - 1,519

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 #14 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 82.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: webdev-html. Source rank: #19. Votes: 324. Organization: google. License: Proprietary.

82.2% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,448 - 1,517

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 #11 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,504
Percentile: 83.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: webdev-react. Source rank: #15. Votes: 1893. Organization: google. License: Proprietary.

83.1% percentile inside its fair comparison set

1,504Raw benchmark valueCI 1,489 - 1,518

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 #9 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 26.7%
Percentile: 63.6%
Last updated: recent
Eligibility: headline eligible

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

63.6% percentile inside its fair comparison set

26.7%Raw benchmark valueCI 18.7% - 34.8%

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 #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 87.6%
Percentile: 95.6%
Last updated: recent
Eligibility: headline eligible

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

95.6% percentile inside its fair comparison set

87.6%Raw benchmark valueCI 85.7% - 89.5%

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 #9 · Source label: google/gemini-3.5-flash

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: Google.

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 #6 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 78.8%
Percentile: 90.7%
Last updated: recent
Eligibility: headline eligible

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

90.7% percentile inside its fair comparison set

78.8%Raw benchmark valueCI 75.2% - 82.4%

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 #3 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 74.2%
Percentile: 92.6%
Last updated: recent
Eligibility: headline eligible

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

92.6% percentile inside its fair comparison set

74.2%Raw benchmark valueCI 72% - 76.4%

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 #12 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 48.7%
Percentile: 77.6%
Last updated: recent
Eligibility: headline eligible

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

77.6% percentile inside its fair comparison set

48.7%Raw benchmark valueCI 39.4% - 57.9%

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 #27 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,506
Percentile: 91.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: coding. Source rank: #35. Votes: 3043. Organization: google. License: Proprietary.

91.9% percentile inside its fair comparison set

1,506Raw benchmark valueCI 1,494 - 1,517

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: coding. Source rank: #17. Votes: 3043. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,479 - 1,502

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 #36 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 51.7%
Percentile: 67.6%
Last updated: archived
Eligibility: headline eligible

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

67.6% percentile inside its fair comparison set

51.7%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 #23 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 78.2%
Percentile: 81.5%
Last updated: archived
Eligibility: headline eligible

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

81.5% percentile inside its fair comparison set

78.2%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 #36 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 45%
Percentile: 73.1%
Last updated: archived
Eligibility: headline eligible

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

73.1% percentile inside its fair comparison set

45%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 #18 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 50%
Percentile: 86.9%
Last updated: archived
Eligibility: headline eligible

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

86.9% percentile inside its fair comparison set

50%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 #66 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 60%
Percentile: 42.6%
Last updated: archived
Eligibility: headline eligible

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

42.6% percentile inside its fair comparison set

60%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 #15 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 80.3%
Percentile: 91.7%
Last updated: archived
Eligibility: headline eligible

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

91.7% percentile inside its fair comparison set

80.3%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 #34 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 76.1%
Percentile: 78.7%
Last updated: archived
Eligibility: headline eligible

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

78.7% percentile inside its fair comparison set

76.1%Raw benchmark value

Reasoning / math / science18 benchmarks88.5%

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 #30 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.1%
Percentile: 92.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `hle`.

92.2% percentile inside its fair comparison set

23.1%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 #46 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 82.8%
Percentile: 88%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `gpqa`.

88% percentile inside its fair comparison set

82.8%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 #38 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 1.4%
Percentile: 88.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `critpt`.

88.1% percentile inside its fair comparison set

1.4%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 #10 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 29%
Percentile: 74.3%
Last updated: recent
Eligibility: headline eligible

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

74.3% percentile inside its fair comparison set

29%Raw benchmark valueCI 20.1% - 37.9%

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 #4 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 92.7%
Percentile: 96.6%
Last updated: recent
Eligibility: headline eligible

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

96.6% percentile inside its fair comparison set

92.7%Raw benchmark valueCI 89.8% - 95.5%

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 #6 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 89.5%
Percentile: 94.4%
Last updated: recent
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

89.5%Raw benchmark valueCI 88.9% - 90.1%

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 #2 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,516
Percentile: 99.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: math. Source rank: #2. Votes: 584. Organization: google. License: Proprietary.

99.7% percentile inside its fair comparison set

1,516Raw benchmark valueCI 1,491 - 1,542

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 #1 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,520
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: math. Source rank: #1. Votes: 584. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,520Raw benchmark valueCI 1,494 - 1,545

Mathematics

LB · Reasoning / math / science · Objective

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

Rank #18 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 88.2%
Percentile: 84.3%
Last updated: archived
Eligibility: headline eligible

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

84.3% percentile inside its fair comparison set

88.2%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 #22 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 82%
Percentile: 80.6%
Last updated: archived
Eligibility: headline eligible

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

80.6% percentile inside its fair comparison set

82%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 #17 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 98%
Percentile: 94.4%
Last updated: archived
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

98%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 #24 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 68%
Percentile: 78.7%
Last updated: archived
Eligibility: headline eligible

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

78.7% percentile inside its fair comparison set

68%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 #20 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 95.1%
Percentile: 89.8%
Last updated: archived
Eligibility: headline eligible

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

89.8% percentile inside its fair comparison set

95.1%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 #4 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 91.9%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

91.9%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 #12 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 80.8%
Percentile: 93.5%
Last updated: archived
Eligibility: headline eligible

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

93.5% percentile inside its fair comparison set

80.8%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 #29 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 77.3%
Percentile: 73.8%
Last updated: archived
Eligibility: headline eligible

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

73.8% percentile inside its fair comparison set

77.3%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 #37 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 96%
Percentile: 73.1%
Last updated: archived
Eligibility: headline eligible

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

73.1% percentile inside its fair comparison set

96%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 #9 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 74%
Percentile: 93.5%
Last updated: archived
Eligibility: headline eligible

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

93.5% percentile inside its fair comparison set

74%Raw benchmark value

Professional reasoning37 benchmarks83.8%

GDPval-AA

AA · Professional reasoning · Rubric

Agentic performance on economically valuable work tasks.

Rank #8 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 1,349
Percentile: 84.8%
Last updated: recent
Eligibility: headline eligible

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

84.8% percentile inside its fair comparison set

1,349Raw benchmark value

APEX-Agents-AA

AA · Professional reasoning · Objective

Long-horizon agentic task completion.

Rank #1 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 47.1%
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `apexAgents`.

100% percentile inside its fair comparison set

47.1%Raw benchmark value

Legal Research Bench

VALS-AI · Professional reasoning · Objective

Applied legal research tasks.

Rank #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 30.8%
Percentile: 66.7%
Last updated: recent
Eligibility: headline eligible

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

66.7% percentile inside its fair comparison set

30.8%Raw benchmark valueCI 24.5% - 37.1%

SkillsBench

VALS-AI · Professional reasoning · Objective

Applied professional skills tasks.

Rank #4 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 52.7%
Percentile: 70%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: skillsbench; provider: Google.

70% percentile inside its fair comparison set

52.7%Raw benchmark valueCI 44% - 61.5%

Public Benefits Bench

VALS-AI · Professional reasoning · Objective

Answering SNAP benefits questions across the public-benefits lifecycle.

Rank #7 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 59.5%
Percentile: 45.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: public-benefits-bench; provider: Google.

45.5% percentile inside its fair comparison set

59.5%Raw benchmark valueCI 57% - 62%

Vals Index

VALS-AI · Professional reasoning · Combined

Weighted model performance across economically relevant Vals tasks.

Rank #6 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 63
Percentile: 80.8%
Last updated: recent
Eligibility: headline eligible

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

80.8% percentile inside its fair comparison set

63Raw benchmark valueCI 60 - 66

Harvey's Legal Agent Benchmark

VALS-AI · Professional reasoning · Objective

Completing legal work with documents, spreadsheets, presentations, and file-system tools.

Rank #8 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 2.5%
Percentile: 46.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: hlab; provider: Google.

46.2% percentile inside its fair comparison set

2.5%Raw benchmark valueCI 0.9% - 4.1%

LegalBench

VALS-AI · Professional reasoning · Objective

Academic legal reasoning tasks.

Rank #29 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 83.6%
Percentile: 68.9%
Last updated: recent
Eligibility: headline eligible

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

68.9% percentile inside its fair comparison set

83.6%Raw benchmark valueCI 81.9% - 85.3%

Finance Agent v2

VALS-AI · Professional reasoning · Objective

Core financial analyst tasks for agentic models.

Rank #1 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 57.9%
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

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

100% percentile inside its fair comparison set

57.9%Raw benchmark valueCI 57.4% - 58.3%

TaxEval v2

VALS-AI · Professional reasoning · Objective

Answer quality on tax questions and responses.

Rank #21 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 74.4%
Percentile: 78%
Last updated: recent
Eligibility: headline eligible

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

78% percentile inside its fair comparison set

74.4%Raw benchmark valueCI 72.7% - 76%

MedCode

VALS-AI · Professional reasoning · Objective

Medical billing support and coding tasks.

Rank #4 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 55.8%
Percentile: 94.1%
Last updated: recent
Eligibility: headline eligible

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

94.1% percentile inside its fair comparison set

55.8%Raw benchmark valueCI 51.7% - 60%

MedScribe

VALS-AI · Professional reasoning · Objective

Administrative documentation support for doctors.

Rank #29 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 76.6%
Percentile: 44%
Last updated: recent
Eligibility: headline eligible

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

44% percentile inside its fair comparison set

76.6%Raw benchmark valueCI 72.8% - 80.3%

Text Arena · Expert

AR · Professional reasoning · Human

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

Rank #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,509
Percentile: 96.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: expert. Source rank: #12. Votes: 1040. Organization: google. License: Proprietary.

96.7% percentile inside its fair comparison set

1,509Raw benchmark valueCI 1,490 - 1,528

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 #28 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,457
Percentile: 91.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_business_and_management_and_financial_operations. Source rank: #39. Votes: 2099. Organization: google. License: Proprietary.

91.5% percentile inside its fair comparison set

1,457Raw benchmark valueCI 1,443 - 1,470

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,448
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_entertainment_and_sports_and_media. Source rank: #16. Votes: 2054. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,448Raw benchmark valueCI 1,434 - 1,462

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

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

Rank #20 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,474
Percentile: 93.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_legal_and_government. Source rank: #26. Votes: 790. Organization: google. License: Proprietary.

93.6% percentile inside its fair comparison set

1,474Raw benchmark valueCI 1,452 - 1,496

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_life_and_physical_and_social_science. Source rank: #16. Votes: 1638. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,480 - 1,510

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

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

Rank #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,509
Percentile: 98.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_mathematical. Source rank: #6. Votes: 644. Organization: google. License: Proprietary.

98.7% percentile inside its fair comparison set

1,509Raw benchmark valueCI 1,484 - 1,533

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

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

Rank #15 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,490
Percentile: 95.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_medicine_and_healthcare. Source rank: #17. Votes: 751. Organization: google. License: Proprietary.

95.3% percentile inside its fair comparison set

1,490Raw benchmark valueCI 1,467 - 1,513

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 #23 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,499
Percentile: 93.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_software_and_it_services. Source rank: #31. Votes: 4221. Organization: google. License: Proprietary.

93.2% percentile inside its fair comparison set

1,499Raw benchmark valueCI 1,489 - 1,509

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_writing_and_literature_and_language. Source rank: #17. Votes: 2335. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,449 - 1,474

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

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

Rank #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,508
Percentile: 98.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: expert. Source rank: #8. Votes: 1040. Organization: google. License: Proprietary.

98.2% percentile inside its fair comparison set

1,508Raw benchmark valueCI 1,489 - 1,527

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 #15 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,456
Percentile: 95.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_business_and_management_and_financial_operations. Source rank: #19. Votes: 2099. Organization: google. License: Proprietary.

95.6% percentile inside its fair comparison set

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

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 #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,454
Percentile: 98.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_entertainment_and_sports_and_media. Source rank: #9. Votes: 2054. Organization: google. License: Proprietary.

98.1% percentile inside its fair comparison set

1,454Raw benchmark valueCI 1,440 - 1,468

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 #8 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,483
Percentile: 97.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_legal_and_government. Source rank: #9. Votes: 790. Organization: google. License: Proprietary.

97.7% percentile inside its fair comparison set

1,483Raw benchmark valueCI 1,461 - 1,504

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 #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,497
Percentile: 98.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_life_and_physical_and_social_science. Source rank: #6. Votes: 1638. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,497Raw benchmark valueCI 1,482 - 1,512

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

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

Rank #2 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,513
Percentile: 99.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_mathematical. Source rank: #3. Votes: 644. Organization: google. License: Proprietary.

99.7% percentile inside its fair comparison set

1,513Raw benchmark valueCI 1,488 - 1,537

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 #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,489
Percentile: 98.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_medicine_and_healthcare. Source rank: #7. Votes: 751. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,489Raw benchmark valueCI 1,466 - 1,511

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 #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 98.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_software_and_it_services. Source rank: #9. Votes: 4221. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,482 - 1,501

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 #9 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,466
Percentile: 97.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_writing_and_literature_and_language. Source rank: #11. Votes: 2335. Organization: google. License: Proprietary.

97.5% percentile inside its fair comparison set

1,466Raw benchmark valueCI 1,454 - 1,479

SAGE

VALS-AI · Professional reasoning · Objective

Student Assessment with Generative Evaluation.

Rank #12 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 49.9%
Percentile: 75.6%
Last updated: recent
Eligibility: headline eligible

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

75.6% percentile inside its fair comparison set

49.9%Raw benchmark valueCI 43.1% - 56.6%

Public Benefits Bench v1

VALS-AI · Professional reasoning · Objective

Answering public-benefits questions across the benefits lifecycle.

Rank #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 58%
Percentile: 66.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: public-benefits-bench-v1; provider: Google.

66.7% percentile inside its fair comparison set

58%Raw benchmark valueCI 55.5% - 60.5%

Data analysis

LB · Professional reasoning · Objective

Structured data manipulation and table reasoning accuracy.

Rank #38 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 64.9%
Percentile: 65.7%
Last updated: archived
Eligibility: headline eligible

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

65.7% percentile inside its fair comparison set

64.9%Raw benchmark value

Overall

LB · Professional reasoning · Objective

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

Rank #14 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75%
Percentile: 88%
Last updated: archived
Eligibility: headline eligible

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

88% percentile inside its fair comparison set

75%Raw benchmark value

Consecutive events

LB · Professional reasoning · Objective

Objective consecutive events score in LiveBench.

Rank #43 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 48%
Percentile: 61.1%
Last updated: archived
Eligibility: headline eligible

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

61.1% percentile inside its fair comparison set

48%Raw benchmark value

Table join

LB · Professional reasoning · Objective

Objective table join score in LiveBench.

Rank #7 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 50.5%
Percentile: 94.4%
Last updated: archived
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

50.5%Raw benchmark value

Table reformat

LB · Professional reasoning · Objective

Objective table reformat score in LiveBench.

Rank #82 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 96.1%
Percentile: 30.6%
Last updated: archived
Eligibility: headline eligible

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

30.6% percentile inside its fair comparison set

96.1%Raw benchmark value

Search / tool use1 benchmark63.1%

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 #115 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 58.8%
Percentile: 63.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `tau2`.

63.1% percentile inside its fair comparison set

58.8%Raw benchmark value

Long context2 benchmarks74.1%

Long Context Reasoning

AA · Long context · Objective

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

Rank #82 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 53.3%
Percentile: 74.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `lcr`.

74.3% percentile inside its fair comparison set

53.3%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 #24 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 64.7%
Percentile: 73.9%
Last updated: recent
Eligibility: headline eligible

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

73.9% percentile inside its fair comparison set

64.7%Raw benchmark valueCI 62.8% - 66.5%

Vision understanding2 benchmarks97.7%

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 #5 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 80.1%
Percentile: 97%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `mmmuPro`.

97% percentile inside its fair comparison set

80.1%Raw benchmark value

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 #2 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 88.3%
Percentile: 98.3%
Last updated: recent
Eligibility: headline eligible

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

98.3% percentile inside its fair comparison set

88.3%Raw benchmark valueCI 86.7% - 89.8%

Document understanding2 benchmarks77.8%

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 #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 63
Percentile: 78.9%
Last updated: recent
Eligibility: headline eligible

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

78.9% percentile inside its fair comparison set

63Raw benchmark valueCI 60 - 66

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 #15 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 68.1%
Percentile: 76.7%
Last updated: recent
Eligibility: headline eligible

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

76.7% percentile inside its fair comparison set

68.1%Raw benchmark valueCI 66.3% - 69.9%

Multilingual10 benchmarks94.1%

Text Arena · Chinese

AR · Multilingual · Human

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

Rank #19 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,507
Percentile: 93.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: chinese. Source rank: #23. Votes: 581. Organization: google. License: Proprietary.

93.9% percentile inside its fair comparison set

1,507Raw benchmark valueCI 1,482 - 1,533

Text Arena · French

AR · Multilingual · Human

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

Rank #39 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,466
Percentile: 82.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: french. Source rank: #50. Votes: 341. Organization: google. License: Proprietary.

82.4% percentile inside its fair comparison set

1,466Raw benchmark valueCI 1,432 - 1,500

Text Arena · Korean

AR · Multilingual · Human

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

Rank #2 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,470
Percentile: 99.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: korean. Source rank: #2. Votes: 209. Organization: google. License: Proprietary.

99.5% percentile inside its fair comparison set

1,470Raw benchmark valueCI 1,430 - 1,511

Text Arena · Russian

AR · Multilingual · Human

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

Rank #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 97.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: russian. Source rank: #9. Votes: 1051. Organization: google. License: Proprietary.

97.9% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,476 - 1,515

Text Arena · Spanish

AR · Multilingual · Human

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

Rank #23 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,460
Percentile: 89.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: spanish. Source rank: #28. Votes: 373. Organization: google. License: Proprietary.

89.7% percentile inside its fair comparison set

1,460Raw benchmark valueCI 1,429 - 1,491

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

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

Rank #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,516
Percentile: 96.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: chinese. Source rank: #14. Votes: 581. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,516Raw benchmark valueCI 1,490 - 1,541

Text Arena · French · No Style Control

AR · Multilingual · Human

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

Rank #34 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,461
Percentile: 84.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: french. Source rank: #40. Votes: 341. Organization: google. License: Proprietary.

84.7% percentile inside its fair comparison set

1,461Raw benchmark valueCI 1,427 - 1,494

Text Arena · Korean · No Style Control

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: korean. Source rank: #1. Votes: 209. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,438 - 1,520

Text Arena · Russian · No Style Control

AR · Multilingual · Human

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

Rank #4 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,496
Percentile: 99%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: russian. Source rank: #4. Votes: 1051. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,496Raw benchmark valueCI 1,477 - 1,515

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

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

Rank #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 97.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: spanish. Source rank: #9. Votes: 373. Organization: google. License: Proprietary.

97.2% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,437 - 1,500

Source links and registry checks

official

vals-ai

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 →

Model profile · Google

Gemini 3.5 Flash

Closed weightsfrontier · registry tag 2026 current

Visible tradeoffs

Reads as visible tradeoffs across the resolved source data.

Visible coverage: 55.5%
Verified coverage: 55.5%
Spread: 92.4%
Last verified: Jun 20, 2026

78%bench fit

textcodevisiondocument9 aliases31 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 / text37 benchmarks84.4%

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 #30 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 35
Percentile: 92.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `intelligenceIndex`.

92.7% percentile inside its fair comparison set

35Raw 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 #12 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 43.1%
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceAccuracy`.

96.3% percentile inside its fair comparison set

43.1%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 #75 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 26.6%
Percentile: 75.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `omniscienceNonHallucination`.

75.2% percentile inside its fair comparison set

26.6%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 #105 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 47.3%
Percentile: 67%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `ifbench`.

67% percentile inside its fair comparison set

47.3%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 #226 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $3.4 /1M tokens
Percentile: 18.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mBlended0To3To1`.

18.5% percentile inside its fair comparison set

$3.4 /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 #220 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $1.5 /1M input tokens
Percentile: 20.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mInputTokens`.

20.7% percentile inside its fair comparison set

$1.5 /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 #234 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $9 /1M output tokens
Percentile: 15.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `price1mOutputTokens`.

15.6% percentile inside its fair comparison set

$9 /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 #36 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 177.8 tokens/s
Percentile: 83.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianOutputTokensPerSecond`.

83.3% percentile inside its fair comparison set

177.8 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 #195 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.61s
Percentile: 7.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstTokenSeconds`.

7.6% percentile inside its fair comparison set

23.61sRaw 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 #142 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.61s
Percentile: 32.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `medianTimeToFirstAnswerTokenSeconds`.

32.9% percentile inside its fair comparison set

23.61sRaw 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 #11 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,476
Percentile: 96.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: overall. Source rank: #13. Votes: 10171. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,476Raw benchmark valueCI 1,470 - 1,483

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 #8 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,464
Percentile: 97.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: creative_writing. Source rank: #10. Votes: 1524. Organization: google. License: Proprietary.

97.8% percentile inside its fair comparison set

1,464Raw benchmark valueCI 1,448 - 1,480

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 #19 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,477
Percentile: 94.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: english. Source rank: #24. Votes: 4948. Organization: google. License: Proprietary.

94.5% percentile inside its fair comparison set

1,477Raw benchmark valueCI 1,468 - 1,486

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,484
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: exclude_ties. Source rank: #15. Votes: 7856. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,484Raw benchmark valueCI 1,476 - 1,493

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 #18 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 94.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts. Source rank: #23. Votes: 6841. Organization: google. License: Proprietary.

94.8% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,484 - 1,499

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 #28 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,484
Percentile: 91.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts_english. Source rank: #36. Votes: 3483. Organization: google. License: Proprietary.

91.7% percentile inside its fair comparison set

1,484Raw benchmark valueCI 1,473 - 1,495

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 #18 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 94.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: instruction_following. Source rank: #25. Votes: 3458. Organization: google. License: Proprietary.

94.8% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,451 - 1,473

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 #18 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,480
Percentile: 94.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: longer_query. Source rank: #24. Votes: 4434. Organization: google. License: Proprietary.

94.4% percentile inside its fair comparison set

1,480Raw benchmark valueCI 1,470 - 1,490

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 #11 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 96.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: multi_turn. Source rank: #13. Votes: 1848. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,487Raw benchmark valueCI 1,473 - 1,502

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 #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,480
Percentile: 98.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: overall. Source rank: #7. Votes: 10171. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,480Raw benchmark valueCI 1,473 - 1,486

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 #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 98.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: creative_writing. Source rank: #8. Votes: 1524. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,453 - 1,485

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 #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 98.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: english. Source rank: #9. Votes: 4948. Organization: google. License: Proprietary.

98.2% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,470 - 1,488

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 #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,486
Percentile: 98.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: exclude_ties. Source rank: #8. Votes: 7856. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,486Raw benchmark valueCI 1,478 - 1,495

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 #9 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,484
Percentile: 97.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts. Source rank: #11. Votes: 6841. Organization: google. License: Proprietary.

97.5% percentile inside its fair comparison set

1,484Raw benchmark valueCI 1,476 - 1,492

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 #17 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 95.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: hard_prompts_english. Source rank: #20. Votes: 3483. Organization: google. License: Proprietary.

95.1% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,467 - 1,489

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,463
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: instruction_following. Source rank: #17. Votes: 3458. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,463Raw benchmark valueCI 1,452 - 1,473

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 #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,481
Percentile: 98%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: longer_query. Source rank: #10. Votes: 4434. Organization: google. License: Proprietary.

98% percentile inside its fair comparison set

1,481Raw benchmark valueCI 1,472 - 1,491

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 #4 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,488
Percentile: 99.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: multi_turn. Source rank: #6. Votes: 1848. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,488Raw benchmark valueCI 1,473 - 1,502

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 #2 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75.6%
Percentile: 99.1%
Last updated: archived
Eligibility: headline eligible

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

99.1% percentile inside its fair comparison set

75.6%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 #6 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 84.6%
Percentile: 95.4%
Last updated: archived
Eligibility: headline eligible

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

95.4% percentile inside its fair comparison set

84.6%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 #3 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75.5%
Percentile: 98.1%
Last updated: archived
Eligibility: headline eligible

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

98.1% percentile inside its fair comparison set

75.5%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 #3 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 69.1%
Percentile: 98.1%
Last updated: archived
Eligibility: headline eligible

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

98.1% percentile inside its fair comparison set

69.1%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 #4 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75.2%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

75.2%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 #4 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 82.6%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

82.6%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 #5 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 100%
Percentile: 100%
Last updated: archived
Eligibility: headline eligible

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

100% percentile inside its fair comparison set

100%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 #7 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 67.8%
Percentile: 94.4%
Last updated: archived
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

67.8%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 #5 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 86%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

86%Raw benchmark value

Coding23 benchmarks83.2%

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 #23 · Source label: Gemini 3.5 Flash (medium)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 39.4%
Percentile: 92.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `terminalbenchHard`.

92.7% percentile inside its fair comparison set

39.4%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 #11 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 48.8%
Percentile: 97.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `scicode`.

97.3% percentile inside its fair comparison set

48.8%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 #5 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 70
Percentile: 94.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `codingIndex`.

94.7% percentile inside its fair comparison set

70Raw 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 #7 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 37
Percentile: 87%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `agenticIndex`.

87% percentile inside its fair comparison set

37Raw 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 #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,506
Percentile: 87.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: overall. Source rank: #14. Votes: 2217. Organization: google. License: Proprietary.

87.7% percentile inside its fair comparison set

1,506Raw benchmark valueCI 1,492 - 1,519

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 #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,506
Percentile: 87.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: webdev. Source rank: #14. Votes: 2217. Organization: google. License: Proprietary.

87.7% percentile inside its fair comparison set

1,506Raw benchmark valueCI 1,492 - 1,519

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 #14 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 82.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: webdev-html. Source rank: #19. Votes: 324. Organization: google. License: Proprietary.

82.2% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,448 - 1,517

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 #11 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,504
Percentile: 83.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: webdev-react. Source rank: #15. Votes: 1893. Organization: google. License: Proprietary.

83.1% percentile inside its fair comparison set

1,504Raw benchmark valueCI 1,489 - 1,518

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 #9 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 26.7%
Percentile: 63.6%
Last updated: recent
Eligibility: headline eligible

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

63.6% percentile inside its fair comparison set

26.7%Raw benchmark valueCI 18.7% - 34.8%

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 #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 87.6%
Percentile: 95.6%
Last updated: recent
Eligibility: headline eligible

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

95.6% percentile inside its fair comparison set

87.6%Raw benchmark valueCI 85.7% - 89.5%

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 #9 · Source label: google/gemini-3.5-flash

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: Google.

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 #6 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 78.8%
Percentile: 90.7%
Last updated: recent
Eligibility: headline eligible

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

90.7% percentile inside its fair comparison set

78.8%Raw benchmark valueCI 75.2% - 82.4%

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 #3 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 74.2%
Percentile: 92.6%
Last updated: recent
Eligibility: headline eligible

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

92.6% percentile inside its fair comparison set

74.2%Raw benchmark valueCI 72% - 76.4%

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 #12 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 48.7%
Percentile: 77.6%
Last updated: recent
Eligibility: headline eligible

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

77.6% percentile inside its fair comparison set

48.7%Raw benchmark valueCI 39.4% - 57.9%

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 #27 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,506
Percentile: 91.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: coding. Source rank: #35. Votes: 3043. Organization: google. License: Proprietary.

91.9% percentile inside its fair comparison set

1,506Raw benchmark valueCI 1,494 - 1,517

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: coding. Source rank: #17. Votes: 3043. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,479 - 1,502

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 #36 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 51.7%
Percentile: 67.6%
Last updated: archived
Eligibility: headline eligible

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

67.6% percentile inside its fair comparison set

51.7%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 #23 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 78.2%
Percentile: 81.5%
Last updated: archived
Eligibility: headline eligible

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

81.5% percentile inside its fair comparison set

78.2%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 #36 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 45%
Percentile: 73.1%
Last updated: archived
Eligibility: headline eligible

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

73.1% percentile inside its fair comparison set

45%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 #18 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 50%
Percentile: 86.9%
Last updated: archived
Eligibility: headline eligible

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

86.9% percentile inside its fair comparison set

50%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 #66 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 60%
Percentile: 42.6%
Last updated: archived
Eligibility: headline eligible

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

42.6% percentile inside its fair comparison set

60%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 #15 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 80.3%
Percentile: 91.7%
Last updated: archived
Eligibility: headline eligible

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

91.7% percentile inside its fair comparison set

80.3%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 #34 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 76.1%
Percentile: 78.7%
Last updated: archived
Eligibility: headline eligible

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

78.7% percentile inside its fair comparison set

76.1%Raw benchmark value

Reasoning / math / science18 benchmarks88.5%

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 #30 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 23.1%
Percentile: 92.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `hle`.

92.2% percentile inside its fair comparison set

23.1%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 #46 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 82.8%
Percentile: 88%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `gpqa`.

88% percentile inside its fair comparison set

82.8%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 #38 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 1.4%
Percentile: 88.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `critpt`.

88.1% percentile inside its fair comparison set

1.4%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 #10 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 29%
Percentile: 74.3%
Last updated: recent
Eligibility: headline eligible

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

74.3% percentile inside its fair comparison set

29%Raw benchmark valueCI 20.1% - 37.9%

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 #4 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 92.7%
Percentile: 96.6%
Last updated: recent
Eligibility: headline eligible

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

96.6% percentile inside its fair comparison set

92.7%Raw benchmark valueCI 89.8% - 95.5%

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 #6 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 89.5%
Percentile: 94.4%
Last updated: recent
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

89.5%Raw benchmark valueCI 88.9% - 90.1%

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 #2 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,516
Percentile: 99.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: math. Source rank: #2. Votes: 584. Organization: google. License: Proprietary.

99.7% percentile inside its fair comparison set

1,516Raw benchmark valueCI 1,491 - 1,542

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 #1 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,520
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: math. Source rank: #1. Votes: 584. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,520Raw benchmark valueCI 1,494 - 1,545

Mathematics

LB · Reasoning / math / science · Objective

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

Rank #18 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 88.2%
Percentile: 84.3%
Last updated: archived
Eligibility: headline eligible

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

84.3% percentile inside its fair comparison set

88.2%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 #22 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 82%
Percentile: 80.6%
Last updated: archived
Eligibility: headline eligible

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

80.6% percentile inside its fair comparison set

82%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 #17 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 98%
Percentile: 94.4%
Last updated: archived
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

98%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 #24 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 68%
Percentile: 78.7%
Last updated: archived
Eligibility: headline eligible

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

78.7% percentile inside its fair comparison set

68%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 #20 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 95.1%
Percentile: 89.8%
Last updated: archived
Eligibility: headline eligible

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

89.8% percentile inside its fair comparison set

95.1%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 #4 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 91.9%
Percentile: 97.2%
Last updated: archived
Eligibility: headline eligible

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

97.2% percentile inside its fair comparison set

91.9%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 #12 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 80.8%
Percentile: 93.5%
Last updated: archived
Eligibility: headline eligible

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

93.5% percentile inside its fair comparison set

80.8%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 #29 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 77.3%
Percentile: 73.8%
Last updated: archived
Eligibility: headline eligible

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

73.8% percentile inside its fair comparison set

77.3%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 #37 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 96%
Percentile: 73.1%
Last updated: archived
Eligibility: headline eligible

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

73.1% percentile inside its fair comparison set

96%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 #9 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 74%
Percentile: 93.5%
Last updated: archived
Eligibility: headline eligible

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

93.5% percentile inside its fair comparison set

74%Raw benchmark value

Professional reasoning37 benchmarks83.8%

GDPval-AA

AA · Professional reasoning · Rubric

Agentic performance on economically valuable work tasks.

Rank #8 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 1,349
Percentile: 84.8%
Last updated: recent
Eligibility: headline eligible

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

84.8% percentile inside its fair comparison set

1,349Raw benchmark value

APEX-Agents-AA

AA · Professional reasoning · Objective

Long-horizon agentic task completion.

Rank #1 · Source label: Gemini 3.5 Flash (high)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 47.1%
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `apexAgents`.

100% percentile inside its fair comparison set

47.1%Raw benchmark value

Legal Research Bench

VALS-AI · Professional reasoning · Objective

Applied legal research tasks.

Rank #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 30.8%
Percentile: 66.7%
Last updated: recent
Eligibility: headline eligible

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

66.7% percentile inside its fair comparison set

30.8%Raw benchmark valueCI 24.5% - 37.1%

SkillsBench

VALS-AI · Professional reasoning · Objective

Applied professional skills tasks.

Rank #4 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 52.7%
Percentile: 70%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: skillsbench; provider: Google.

70% percentile inside its fair comparison set

52.7%Raw benchmark valueCI 44% - 61.5%

Public Benefits Bench

VALS-AI · Professional reasoning · Objective

Answering SNAP benefits questions across the public-benefits lifecycle.

Rank #7 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 59.5%
Percentile: 45.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: public-benefits-bench; provider: Google.

45.5% percentile inside its fair comparison set

59.5%Raw benchmark valueCI 57% - 62%

Vals Index

VALS-AI · Professional reasoning · Combined

Weighted model performance across economically relevant Vals tasks.

Rank #6 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 63
Percentile: 80.8%
Last updated: recent
Eligibility: headline eligible

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

80.8% percentile inside its fair comparison set

63Raw benchmark valueCI 60 - 66

Harvey's Legal Agent Benchmark

VALS-AI · Professional reasoning · Objective

Completing legal work with documents, spreadsheets, presentations, and file-system tools.

Rank #8 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 2.5%
Percentile: 46.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: hlab; provider: Google.

46.2% percentile inside its fair comparison set

2.5%Raw benchmark valueCI 0.9% - 4.1%

LegalBench

VALS-AI · Professional reasoning · Objective

Academic legal reasoning tasks.

Rank #29 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 83.6%
Percentile: 68.9%
Last updated: recent
Eligibility: headline eligible

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

68.9% percentile inside its fair comparison set

83.6%Raw benchmark valueCI 81.9% - 85.3%

Finance Agent v2

VALS-AI · Professional reasoning · Objective

Core financial analyst tasks for agentic models.

Rank #1 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 57.9%
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

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

100% percentile inside its fair comparison set

57.9%Raw benchmark valueCI 57.4% - 58.3%

TaxEval v2

VALS-AI · Professional reasoning · Objective

Answer quality on tax questions and responses.

Rank #21 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 74.4%
Percentile: 78%
Last updated: recent
Eligibility: headline eligible

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

78% percentile inside its fair comparison set

74.4%Raw benchmark valueCI 72.7% - 76%

MedCode

VALS-AI · Professional reasoning · Objective

Medical billing support and coding tasks.

Rank #4 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 55.8%
Percentile: 94.1%
Last updated: recent
Eligibility: headline eligible

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

94.1% percentile inside its fair comparison set

55.8%Raw benchmark valueCI 51.7% - 60%

MedScribe

VALS-AI · Professional reasoning · Objective

Administrative documentation support for doctors.

Rank #29 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 76.6%
Percentile: 44%
Last updated: recent
Eligibility: headline eligible

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

44% percentile inside its fair comparison set

76.6%Raw benchmark valueCI 72.8% - 80.3%

Text Arena · Expert

AR · Professional reasoning · Human

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

Rank #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,509
Percentile: 96.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: expert. Source rank: #12. Votes: 1040. Organization: google. License: Proprietary.

96.7% percentile inside its fair comparison set

1,509Raw benchmark valueCI 1,490 - 1,528

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 #28 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,457
Percentile: 91.5%
Last updated: recent
Eligibility: headline eligible

91.5% percentile inside its fair comparison set

1,457Raw benchmark valueCI 1,443 - 1,470

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,448
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_entertainment_and_sports_and_media. Source rank: #16. Votes: 2054. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,448Raw benchmark valueCI 1,434 - 1,462

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

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

Rank #20 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,474
Percentile: 93.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_legal_and_government. Source rank: #26. Votes: 790. Organization: google. License: Proprietary.

93.6% percentile inside its fair comparison set

1,474Raw benchmark valueCI 1,452 - 1,496

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_life_and_physical_and_social_science. Source rank: #16. Votes: 1638. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,480 - 1,510

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

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

Rank #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,509
Percentile: 98.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_mathematical. Source rank: #6. Votes: 644. Organization: google. License: Proprietary.

98.7% percentile inside its fair comparison set

1,509Raw benchmark valueCI 1,484 - 1,533

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

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

Rank #15 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,490
Percentile: 95.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_medicine_and_healthcare. Source rank: #17. Votes: 751. Organization: google. License: Proprietary.

95.3% percentile inside its fair comparison set

1,490Raw benchmark valueCI 1,467 - 1,513

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 #23 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,499
Percentile: 93.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_software_and_it_services. Source rank: #31. Votes: 4221. Organization: google. License: Proprietary.

93.2% percentile inside its fair comparison set

1,499Raw benchmark valueCI 1,489 - 1,509

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 #13 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 96.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_writing_and_literature_and_language. Source rank: #17. Votes: 2335. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,449 - 1,474

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

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

Rank #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,508
Percentile: 98.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: expert. Source rank: #8. Votes: 1040. Organization: google. License: Proprietary.

98.2% percentile inside its fair comparison set

1,508Raw benchmark valueCI 1,489 - 1,527

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 #15 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,456
Percentile: 95.6%
Last updated: recent
Eligibility: headline eligible

95.6% percentile inside its fair comparison set

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

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 #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,454
Percentile: 98.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_entertainment_and_sports_and_media. Source rank: #9. Votes: 2054. Organization: google. License: Proprietary.

98.1% percentile inside its fair comparison set

1,454Raw benchmark valueCI 1,440 - 1,468

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 #8 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,483
Percentile: 97.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_legal_and_government. Source rank: #9. Votes: 790. Organization: google. License: Proprietary.

97.7% percentile inside its fair comparison set

1,483Raw benchmark valueCI 1,461 - 1,504

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 #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,497
Percentile: 98.8%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_life_and_physical_and_social_science. Source rank: #6. Votes: 1638. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,497Raw benchmark valueCI 1,482 - 1,512

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

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

Rank #2 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,513
Percentile: 99.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_mathematical. Source rank: #3. Votes: 644. Organization: google. License: Proprietary.

99.7% percentile inside its fair comparison set

1,513Raw benchmark valueCI 1,488 - 1,537

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 #5 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,489
Percentile: 98.6%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_medicine_and_healthcare. Source rank: #7. Votes: 751. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,489Raw benchmark valueCI 1,466 - 1,511

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 #6 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 98.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_software_and_it_services. Source rank: #9. Votes: 4221. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,482 - 1,501

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 #9 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,466
Percentile: 97.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: industry_writing_and_literature_and_language. Source rank: #11. Votes: 2335. Organization: google. License: Proprietary.

97.5% percentile inside its fair comparison set

1,466Raw benchmark valueCI 1,454 - 1,479

SAGE

VALS-AI · Professional reasoning · Objective

Student Assessment with Generative Evaluation.

Rank #12 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 49.9%
Percentile: 75.6%
Last updated: recent
Eligibility: headline eligible

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

75.6% percentile inside its fair comparison set

49.9%Raw benchmark valueCI 43.1% - 56.6%

Public Benefits Bench v1

VALS-AI · Professional reasoning · Objective

Answering public-benefits questions across the benefits lifecycle.

Rank #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 58%
Percentile: 66.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Vals AI BenchmarkView overall scores. Vals slug: public-benefits-bench-v1; provider: Google.

66.7% percentile inside its fair comparison set

58%Raw benchmark valueCI 55.5% - 60.5%

Data analysis

LB · Professional reasoning · Objective

Structured data manipulation and table reasoning accuracy.

Rank #38 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 64.9%
Percentile: 65.7%
Last updated: archived
Eligibility: headline eligible

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

65.7% percentile inside its fair comparison set

64.9%Raw benchmark value

Overall

LB · Professional reasoning · Objective

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

Rank #14 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 75%
Percentile: 88%
Last updated: archived
Eligibility: headline eligible

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

88% percentile inside its fair comparison set

75%Raw benchmark value

Consecutive events

LB · Professional reasoning · Objective

Objective consecutive events score in LiveBench.

Rank #43 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 48%
Percentile: 61.1%
Last updated: archived
Eligibility: headline eligible

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

61.1% percentile inside its fair comparison set

48%Raw benchmark value

Table join

LB · Professional reasoning · Objective

Objective table join score in LiveBench.

Rank #7 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 50.5%
Percentile: 94.4%
Last updated: archived
Eligibility: headline eligible

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

94.4% percentile inside its fair comparison set

50.5%Raw benchmark value

Table reformat

LB · Professional reasoning · Objective

Objective table reformat score in LiveBench.

Rank #82 · Source label: gemini-3.5-flash-high

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: LiveBench
Raw value: 96.1%
Percentile: 30.6%
Last updated: archived
Eligibility: headline eligible

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

30.6% percentile inside its fair comparison set

96.1%Raw benchmark value

Search / tool use1 benchmark63.1%

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 #115 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 58.8%
Percentile: 63.1%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `tau2`.

63.1% percentile inside its fair comparison set

58.8%Raw benchmark value

Long context2 benchmarks74.1%

Long Context Reasoning

AA · Long context · Objective

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

Rank #82 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 53.3%
Percentile: 74.3%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `lcr`.

74.3% percentile inside its fair comparison set

53.3%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 #24 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 64.7%
Percentile: 73.9%
Last updated: recent
Eligibility: headline eligible

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

73.9% percentile inside its fair comparison set

64.7%Raw benchmark valueCI 62.8% - 66.5%

Vision understanding2 benchmarks97.7%

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 #5 · Source label: Gemini 3.5 Flash (minimal)

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 80.1%
Percentile: 97%
Last updated: recent
Eligibility: headline eligible

Parsed from Artificial Analysis public leaderboard field `mmmuPro`.

97% percentile inside its fair comparison set

80.1%Raw benchmark value

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 #2 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 88.3%
Percentile: 98.3%
Last updated: recent
Eligibility: headline eligible

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

98.3% percentile inside its fair comparison set

88.3%Raw benchmark valueCI 86.7% - 89.8%

Document understanding2 benchmarks77.8%

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 #5 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 63
Percentile: 78.9%
Last updated: recent
Eligibility: headline eligible

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

78.9% percentile inside its fair comparison set

63Raw benchmark valueCI 60 - 66

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 #15 · Source label: google/gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 68.1%
Percentile: 76.7%
Last updated: recent
Eligibility: headline eligible

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

76.7% percentile inside its fair comparison set

68.1%Raw benchmark valueCI 66.3% - 69.9%

Multilingual10 benchmarks94.1%

Text Arena · Chinese

AR · Multilingual · Human

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

Rank #19 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,507
Percentile: 93.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: chinese. Source rank: #23. Votes: 581. Organization: google. License: Proprietary.

93.9% percentile inside its fair comparison set

1,507Raw benchmark valueCI 1,482 - 1,533

Text Arena · French

AR · Multilingual · Human

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

Rank #39 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,466
Percentile: 82.4%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: french. Source rank: #50. Votes: 341. Organization: google. License: Proprietary.

82.4% percentile inside its fair comparison set

1,466Raw benchmark valueCI 1,432 - 1,500

Text Arena · Korean

AR · Multilingual · Human

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

Rank #2 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,470
Percentile: 99.5%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: korean. Source rank: #2. Votes: 209. Organization: google. License: Proprietary.

99.5% percentile inside its fair comparison set

1,470Raw benchmark valueCI 1,430 - 1,511

Text Arena · Russian

AR · Multilingual · Human

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

Rank #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 97.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: russian. Source rank: #9. Votes: 1051. Organization: google. License: Proprietary.

97.9% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,476 - 1,515

Text Arena · Spanish

AR · Multilingual · Human

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

Rank #23 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,460
Percentile: 89.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: spanish. Source rank: #28. Votes: 373. Organization: google. License: Proprietary.

89.7% percentile inside its fair comparison set

1,460Raw benchmark valueCI 1,429 - 1,491

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

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

Rank #10 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,516
Percentile: 96.9%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: chinese. Source rank: #14. Votes: 581. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,516Raw benchmark valueCI 1,490 - 1,541

Text Arena · French · No Style Control

AR · Multilingual · Human

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

Rank #34 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,461
Percentile: 84.7%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: french. Source rank: #40. Votes: 341. Organization: google. License: Proprietary.

84.7% percentile inside its fair comparison set

1,461Raw benchmark valueCI 1,427 - 1,494

Text Arena · Korean · No Style Control

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 100%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: korean. Source rank: #1. Votes: 209. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,438 - 1,520

Text Arena · Russian · No Style Control

AR · Multilingual · Human

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

Rank #4 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,496
Percentile: 99%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: russian. Source rank: #4. Votes: 1051. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,496Raw benchmark valueCI 1,477 - 1,515

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

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

Rank #7 · Source label: gemini-3.5-flash

verified runtimeexact alias

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 97.2%
Last updated: recent
Eligibility: headline eligible

Parsed from Arena leaderboard dataset row `gemini-3.5-flash`. Category: spanish. Source rank: #9. Votes: 373. Organization: google. License: Proprietary.

97.2% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,437 - 1,500