Model profile · Google

Gemini 3 Pro Preview

Closed weightsfrontier · registry tag 2026 preview

Thin verified coverage

Reads as thin verified coverage across the resolved source data.

Visible coverage: 34.2%
Verified coverage: 34.2%
Spread: n/a
Last verified: Jun 20, 2026

textcodevisiondocumentaudiosearch6 aliases48 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.

Thin verified coverageThis model currently reads as thin verified coverage across the resolved source data.

Chat / text26 benchmarks80.5%

Intelligence Index

AA · Chat / text · Combined

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

Rank #39 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 33
Percentile: 90.4%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `intelligenceIndex`.

90.4% percentile inside its fair comparison set

33Raw 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 #4 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 47.1%
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `omniscienceAccuracy`.

99% percentile inside its fair comparison set

47.1%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 #229 · Source label: Gemini 3 Pro Preview (high)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 9.1%
Percentile: 23.5%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `omniscienceNonHallucination`.

23.5% percentile inside its fair comparison set

9.1%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 #94 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 49.7%
Percentile: 70.5%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `ifbench`.

70.5% percentile inside its fair comparison set

49.7%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 #242 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $4.5 /1M tokens
Percentile: 13%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `price1mBlended0To3To1`.

13% percentile inside its fair comparison set

$4.5 /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 #237 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $2 /1M input tokens
Percentile: 17.8%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `price1mInputTokens`.

17.8% percentile inside its fair comparison set

$2 /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 #242 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $12 /1M output tokens
Percentile: 13%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `price1mOutputTokens`.

13% percentile inside its fair comparison set

$12 /1M output tokensRaw benchmark value

Openness Index

AA · Chat / text · Combined

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

Rank #185 · Source label: Gemini 3 Pro Preview (high)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

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

7.5% percentile inside its fair comparison set

6Raw benchmark value

Text Arena

AR · Chat / text · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #8. Votes: 41314. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,486Raw benchmark valueCI 1,482 - 1,490

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,485
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #4. Votes: 6288. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,485Raw benchmark valueCI 1,476 - 1,493

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 #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,489
Percentile: 98.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #8. Votes: 18965. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,489Raw benchmark valueCI 1,484 - 1,494

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,501
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: exclude_ties. Source rank: #6. Votes: 29671. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,501Raw benchmark valueCI 1,496 - 1,507

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 #7 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,504
Percentile: 98.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts. Source rank: #9. Votes: 22462. Organization: google. License: Proprietary.

98.2% percentile inside its fair comparison set

1,504Raw benchmark valueCI 1,499 - 1,509

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,502
Percentile: 96.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts_english. Source rank: #14. Votes: 10730. Organization: google. License: Proprietary.

96.6% percentile inside its fair comparison set

1,502Raw benchmark valueCI 1,496 - 1,509

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 #11 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,474
Percentile: 96.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: instruction_following. Source rank: #14. Votes: 11193. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,474Raw benchmark valueCI 1,468 - 1,481

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 #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,492
Percentile: 97.4%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: longer_query. Source rank: #13. Votes: 10490. Organization: google. License: Proprietary.

97.4% percentile inside its fair comparison set

1,492Raw benchmark valueCI 1,486 - 1,499

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 #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 98.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: multi_turn. Source rank: #8. Votes: 6769. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,487 - 1,503

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 #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 98.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #8. Votes: 41314. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,476 - 1,483

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #5. Votes: 6288. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,474 - 1,490

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #6. Votes: 18965. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,474 - 1,485

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: exclude_ties. Source rank: #5. Votes: 29671. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,486 - 1,496

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 #11 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 96.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts. Source rank: #14. Votes: 22462. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,477 - 1,487

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 #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 95.4%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts_english. Source rank: #19. Votes: 10730. Organization: google. License: Proprietary.

95.4% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,472 - 1,484

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 #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,459
Percentile: 95.4%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: instruction_following. Source rank: #21. Votes: 11193. Organization: google. License: Proprietary.

95.4% percentile inside its fair comparison set

1,459Raw benchmark valueCI 1,453 - 1,465

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 #17 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,473
Percentile: 94.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: longer_query. Source rank: #22. Votes: 10490. Organization: google. License: Proprietary.

94.7% percentile inside its fair comparison set

1,473Raw benchmark valueCI 1,467 - 1,480

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 #7 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,483
Percentile: 98.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: multi_turn. Source rank: #9. Votes: 6769. Organization: google. License: Proprietary.

98.1% percentile inside its fair comparison set

1,483Raw benchmark valueCI 1,475 - 1,490

Coding16 benchmarks67.7%

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 #46 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 34.1%
Percentile: 85.1%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `terminalbenchHard`.

85.1% percentile inside its fair comparison set

34.1%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 #9 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 49.9%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `scicode`.

97.8% percentile inside its fair comparison set

49.9%Raw 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 #23 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,439
Percentile: 69.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #29. Votes: 17158. Organization: google. License: Proprietary.

69.9% percentile inside its fair comparison set

1,439Raw benchmark valueCI 1,431 - 1,446

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 #23 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,439
Percentile: 69.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: webdev. Source rank: #29. Votes: 17158. Organization: google. License: Proprietary.

69.9% percentile inside its fair comparison set

1,439Raw benchmark valueCI 1,431 - 1,446

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 #20 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 74%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: webdev-html. Source rank: #26. Votes: 13755. Organization: google. License: Proprietary.

74% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,453 - 1,472

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 #28 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,402
Percentile: 54.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: webdev-react. Source rank: #35. Votes: 3394. Organization: google. License: Proprietary.

54.2% percentile inside its fair comparison set

1,402Raw benchmark valueCI 1,391 - 1,413

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 #12 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 86.4%
Percentile: 87.8%
Last updated: recent
Eligibility: preview_model

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

87.8% percentile inside its fair comparison set

86.4%Raw benchmark valueCI 84.5% - 88.3%

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 #15 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 76.4%
Percentile: 75.9%
Last updated: recent
Eligibility: preview_model

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

75.9% percentile inside its fair comparison set

76.4%Raw benchmark valueCI 72.7% - 80.1%

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 #36 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 14.3%
Percentile: 28.6%
Last updated: recent
Eligibility: preview_model

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

28.6% percentile inside its fair comparison set

14.3%Raw benchmark valueCI 8.3% - 20.3%

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 #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,519
Percentile: 95.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: coding. Source rank: #20. Votes: 8573. Organization: google. License: Proprietary.

95.3% percentile inside its fair comparison set

1,519Raw benchmark valueCI 1,512 - 1,526

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 #20 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,483
Percentile: 94.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: coding. Source rank: #25. Votes: 8573. Organization: google. License: Proprietary.

94.1% percentile inside its fair comparison set

1,483Raw benchmark valueCI 1,476 - 1,490

IOI

VALS-AI · Coding · Objective

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

Rank #8 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 38.8%
Percentile: 84.1%
Last updated: recent
Eligibility: preview_model

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

84.1% percentile inside its fair comparison set

38.8%Raw benchmark valueCI 16.3% - 61.4%

Code Arena · Image To Webdev

AR · Coding · Human

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

Rank #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,453
Percentile: 52.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: image_to_webdev. Source rank: #12. Votes: 1091. Organization: google. License: Proprietary.

52.9% percentile inside its fair comparison set

1,453Raw benchmark valueCI 1,433 - 1,473

Terminal-Bench 2.0

OFF · Coding · Objective

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

Rank #7

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 56.9%
Percentile: 0%
Last updated: recent
Eligibility: preview_model

Verified against the current Gemini Pro public model page. Page did not expose a visible page date on April 29, 2026. Uses the published Terminal-Bench 2.0 figure for Gemini 3 Pro Preview. Checked 2026-04-29. Verification: manual_public_page_verification.

0% percentile inside its fair comparison set

56.9%Raw benchmark value

SWE-Bench Verified

OFF · Coding · Objective

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

Rank #5

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 76.2%
Percentile: 20%
Last updated: recent
Eligibility: preview_model

Verified against the current Gemini Pro public model page. Page did not expose a visible page date on April 29, 2026. Uses the published SWE-Bench Verified figure for Gemini 3 Pro Preview. Checked 2026-04-29. Verification: manual_public_page_verification.

20% percentile inside its fair comparison set

76.2%Raw benchmark value

Terminal-Bench 2.0

TERMINAL-BENCH · Coding · Objective

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

Rank #3

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Terminal-Bench
Raw value: 69.4%
Percentile: 93.3%
Last updated: archived
Eligibility: preview_model

Parsed from the public Terminal-Bench 2.0 verified leaderboard. Collapse policy: highest verified score per canonical model. Selected agent: Ante (unknown). Display model: Gemini 3 Pro. Integration method: API. Agent URL: https://antigma.ai. Reported stderr: 1.066 percentage points.

93.3% percentile inside its fair comparison set

69.4%Raw benchmark value

Reasoning / math / science11 benchmarks77%

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 #18 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 27.6%
Percentile: 95.4%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `hle`.

95.4% percentile inside its fair comparison set

27.6%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 #12 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 88.7%
Percentile: 97.1%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `gpqa`.

97.1% percentile inside its fair comparison set

88.7%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 #108 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

Parsed from Artificial Analysis public leaderboard field `critpt`.

65.2% percentile inside its fair comparison set

0%Raw benchmark value

ProofBench

VALS-AI · Reasoning / math / science · Objective

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

Rank #16 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 20%
Percentile: 57.1%
Last updated: recent
Eligibility: preview_model

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

57.1% percentile inside its fair comparison set

20%Raw benchmark valueCI 12.2% - 27.8%

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 #7 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 91.7%
Percentile: 93.3%
Last updated: recent
Eligibility: preview_model

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

93.3% percentile inside its fair comparison set

91.7%Raw benchmark valueCI 88.9% - 94.4%

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 #3 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 90.1%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

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

97.8% percentile inside its fair comparison set

90.1%Raw benchmark valueCI 89.5% - 90.7%

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 #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 96.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: math. Source rank: #17. Votes: 2652. Organization: google. License: Proprietary.

96.2% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,466 - 1,489

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,476
Percentile: 96.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: math. Source rank: #15. Votes: 2652. Organization: google. License: Proprietary.

96.5% percentile inside its fair comparison set

1,476Raw benchmark valueCI 1,465 - 1,487

TutorBench

SL · Reasoning / math / science · Rubric

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

Rank #7

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 53.7%
Percentile: 40%
Last updated: recent
Eligibility: preview_model

40% percentile inside its fair comparison set

53.7%Raw benchmark value

MultiNRC

SL · Reasoning / math / science · Rubric

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

Rank #3

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 59%
Percentile: 80%
Last updated: recent
Eligibility: preview_model

80% percentile inside its fair comparison set

59%Raw benchmark value

Humanity's Last Exam

OFF · Reasoning / math / science · Objective

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

Rank #6

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 37.5%
Percentile: 28.6%
Last updated: recent
Eligibility: preview_model

Verified against the current Gemini Pro public model page. Page did not expose a visible page date on April 29, 2026. Uses the published no-tools Humanity's Last Exam figure for Gemini 3 Pro Preview. Checked 2026-04-29. Verification: manual_public_page_verification.

28.6% percentile inside its fair comparison set

37.5%Raw benchmark value

Professional reasoning25 benchmarks85.7%

LegalBench

VALS-AI · Professional reasoning · Objective

Academic legal reasoning tasks.

Rank #3 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 87%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

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

97.8% percentile inside its fair comparison set

87%Raw benchmark valueCI 86.3% - 87.7%

TaxEval v2

VALS-AI · Professional reasoning · Objective

Answer quality on tax questions and responses.

Rank #42 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 72.6%
Percentile: 54.9%
Last updated: recent
Eligibility: preview_model

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

54.9% percentile inside its fair comparison set

72.6%Raw benchmark valueCI 70.9% - 74.3%

MedCode

VALS-AI · Professional reasoning · Objective

Medical billing support and coding tasks.

Rank #8 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 52.2%
Percentile: 86.3%
Last updated: recent
Eligibility: preview_model

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

86.3% percentile inside its fair comparison set

52.2%Raw benchmark valueCI 48.1% - 56.3%

MedScribe

VALS-AI · Professional reasoning · Objective

Administrative documentation support for doctors.

Rank #43 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 72%
Percentile: 16%
Last updated: recent
Eligibility: preview_model

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

16% percentile inside its fair comparison set

72%Raw benchmark valueCI 68.3% - 75.8%

Text Arena · Expert

AR · Professional reasoning · Human

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

Rank #14 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,503
Percentile: 95.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: expert. Source rank: #19. Votes: 2532. Organization: google. License: Proprietary.

95.3% percentile inside its fair comparison set

1,503Raw benchmark valueCI 1,491 - 1,515

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,475
Percentile: 96.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_business_and_management_and_financial_operations. Source rank: #16. Votes: 7578. Organization: google. License: Proprietary.

96.5% percentile inside its fair comparison set

1,475Raw benchmark valueCI 1,468 - 1,483

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,468
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_entertainment_and_sports_and_media. Source rank: #6. Votes: 7675. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,468Raw benchmark valueCI 1,461 - 1,476

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,502
Percentile: 98.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_legal_and_government. Source rank: #7. Votes: 2904. Organization: google. License: Proprietary.

98.3% percentile inside its fair comparison set

1,502Raw benchmark valueCI 1,491 - 1,513

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 #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,499
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_life_and_physical_and_social_science. Source rank: #10. Votes: 6719. Organization: google. License: Proprietary.

97.8% percentile inside its fair comparison set

1,499Raw benchmark valueCI 1,491 - 1,506

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

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

Rank #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 95.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_mathematical. Source rank: #19. Votes: 1941. Organization: google. License: Proprietary.

95.1% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,468 - 1,495

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

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

Rank #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,508
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_medicine_and_healthcare. Source rank: #6. Votes: 2532. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,508Raw benchmark valueCI 1,496 - 1,521

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,511
Percentile: 96.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_software_and_it_services. Source rank: #15. Votes: 14254. Organization: google. License: Proprietary.

96.6% percentile inside its fair comparison set

1,511Raw benchmark valueCI 1,505 - 1,517

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,481
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_writing_and_literature_and_language. Source rank: #6. Votes: 9248. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,481Raw benchmark valueCI 1,474 - 1,488

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

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

Rank #18 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,480
Percentile: 93.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: expert. Source rank: #24. Votes: 2532. Organization: google. License: Proprietary.

93.8% percentile inside its fair comparison set

1,480Raw benchmark valueCI 1,468 - 1,492

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 #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,460
Percentile: 96.2%
Last updated: recent
Eligibility: preview_model

96.2% percentile inside its fair comparison set

1,460Raw benchmark valueCI 1,453 - 1,468

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 #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,464
Percentile: 98.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_entertainment_and_sports_and_media. Source rank: #7. Votes: 7675. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,464Raw benchmark valueCI 1,456 - 1,471

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,493
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_legal_and_government. Source rank: #5. Votes: 2904. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,493Raw benchmark valueCI 1,482 - 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 #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,486
Percentile: 97.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_life_and_physical_and_social_science. Source rank: #11. Votes: 6719. Organization: google. License: Proprietary.

97.5% percentile inside its fair comparison set

1,486Raw benchmark valueCI 1,478 - 1,494

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

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

Rank #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,475
Percentile: 95.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_mathematical. Source rank: #19. Votes: 1941. Organization: google. License: Proprietary.

95.1% percentile inside its fair comparison set

1,475Raw benchmark valueCI 1,462 - 1,488

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,492
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_medicine_and_healthcare. Source rank: #5. Votes: 2532. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,492Raw benchmark valueCI 1,480 - 1,504

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 #15 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 95.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_software_and_it_services. Source rank: #18. Votes: 14254. Organization: google. License: Proprietary.

95.7% percentile inside its fair comparison set

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

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 #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 98.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_writing_and_literature_and_language. Source rank: #7. Votes: 9248. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,471 - 1,485

SAGE

VALS-AI · Professional reasoning · Objective

Student Assessment with Generative Evaluation.

Rank #17 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 47.6%
Percentile: 64.4%
Last updated: recent
Eligibility: preview_model

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

64.4% percentile inside its fair comparison set

47.6%Raw benchmark valueCI 41% - 54.2%

PRBench Legal

SL · Professional reasoning · Rubric

Applied legal reasoning on professional-domain tasks.

Rank #11

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 40.6%
Percentile: 16.7%
Last updated: recent
Eligibility: preview_model

16.7% percentile inside its fair comparison set

40.6%Raw benchmark value

Poker Agent

VALS-AI · Professional reasoning · Objective

Agent profit in poker-style strategic play.

Rank #8 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 1,078.9 score
Percentile: 56.3%
Last updated: archived
Eligibility: preview_model

Parsed from Vals AI BenchmarkView overall scores. Vals slug: poker_agent; provider: unknown.

56.3% percentile inside its fair comparison set

1,078.9 scoreRaw benchmark valueCI 1,078.9 score - 1,078.9 score

Search / tool use4 benchmarks56.2%

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 #100 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 68.1%
Percentile: 68%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `tau2`.

68% percentile inside its fair comparison set

68.1%Raw benchmark value

Search Arena

AR · Search / tool use · Human

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

Rank #7 · Source label: gemini-3-pro-grounding

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,201
Percentile: 80%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro-grounding`. Category: overall. Source rank: #7. Votes: 37327. Organization: google. License: Proprietary.

80% percentile inside its fair comparison set

1,201Raw benchmark valueCI 1,195 - 1,206

Search Arena · No Style Control

AR · Search / tool use · Human

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

Rank #8 · Source label: gemini-3-pro-grounding

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,207
Percentile: 76.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro-grounding`. Category: overall. Source rank: #8. Votes: 37327. Organization: google. License: Proprietary.

76.7% percentile inside its fair comparison set

1,207Raw benchmark valueCI 1,201 - 1,212

BrowseComp

OFF · Search / tool use · Objective

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

Rank #7

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 59.2%
Percentile: 0%
Last updated: recent
Eligibility: preview_model

Verified against the current Gemini Pro public model page. Page did not expose a visible page date on April 29, 2026. Uses the published BrowseComp figure for Gemini 3 Pro Preview. Checked 2026-04-29. Verification: manual_public_page_verification.

0% percentile inside its fair comparison set

59.2%Raw benchmark value

Long context3 benchmarks54.5%

Long Context Reasoning

AA · Long context · Objective

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

Rank #16 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 67.3%
Percentile: 95.2%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `lcr`.

95.2% percentile inside its fair comparison set

67.3%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 #30 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 63.7%
Percentile: 68.2%
Last updated: recent
Eligibility: preview_model

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

68.2% percentile inside its fair comparison set

63.7%Raw benchmark valueCI 61.8% - 65.5%

MRCR v2

OFF · Long context · Objective

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

Rank #2

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 77%
Percentile: 0%
Last updated: recent
Eligibility: preview_model

Verified against the current Gemini Pro public model page. Page did not expose a visible page date on April 29, 2026. Uses the published MRCR v2 8-needle 128k average for Gemini 3 Pro Preview. Checked 2026-04-29. Verification: manual_public_page_verification.

0% percentile inside its fair comparison set

77%Raw benchmark value

Vision understanding24 benchmarks92.8%

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 #4 · Source label: Gemini 3 Pro Preview (high)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 80.2%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `mmmuPro`.

97.8% percentile inside its fair comparison set

80.2%Raw benchmark value

Vision Arena

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,290
Percentile: 96.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #7. Votes: 13200. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,290Raw benchmark valueCI 1,282 - 1,297

Vision Arena · Captioning

AR · Vision understanding · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,273
Percentile: 96.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: captioning. Source rank: #2. Votes: 222. Organization: google. License: Proprietary.

96.2% percentile inside its fair comparison set

1,273Raw benchmark valueCI 1,238 - 1,309

Vision Arena · Creative Writing Vision

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,333
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing_vision. Source rank: #1. Votes: 1039. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,333Raw benchmark valueCI 1,314 - 1,352

Vision Arena · Diagram

AR · Vision understanding · Human

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

Rank #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,301
Percentile: 84.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: diagram. Source rank: #16. Votes: 2880. Organization: google. License: Proprietary.

84.3% percentile inside its fair comparison set

1,301Raw benchmark valueCI 1,290 - 1,313

Vision Arena · English

AR · Vision understanding · Human

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

Rank #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,276
Percentile: 89%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #15. Votes: 5347. Organization: google. License: Proprietary.

89% percentile inside its fair comparison set

1,276Raw benchmark valueCI 1,265 - 1,286

Vision Arena · Entity Recognition

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,305
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: entity_recognition. Source rank: #1. Votes: 243. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,305Raw benchmark valueCI 1,269 - 1,340

Vision Arena · Homework

AR · Vision understanding · Human

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

Rank #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,311
Percentile: 88.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: homework. Source rank: #12. Votes: 1795. Organization: google. License: Proprietary.

88.2% percentile inside its fair comparison set

1,311Raw benchmark valueCI 1,296 - 1,325

Vision Arena · Humor

AR · Vision understanding · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,316
Percentile: 98%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: humor. Source rank: #2. Votes: 564. Organization: google. License: Proprietary.

98% percentile inside its fair comparison set

1,316Raw benchmark valueCI 1,291 - 1,341

Vision Arena · Ocr

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,303
Percentile: 94.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: ocr. Source rank: #7. Votes: 8110. Organization: google. License: Proprietary.

94.3% percentile inside its fair comparison set

1,303Raw benchmark valueCI 1,295 - 1,312

Vision Arena · No Style Control

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,305
Percentile: 96.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #7. Votes: 13200. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,305Raw benchmark valueCI 1,298 - 1,313

Vision Arena · Captioning · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,321
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: captioning. Source rank: #1. Votes: 222. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,321Raw benchmark valueCI 1,287 - 1,355

Vision Arena · Creative Writing Vision · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,350
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing_vision. Source rank: #1. Votes: 1039. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,350Raw benchmark valueCI 1,331 - 1,369

Vision Arena · Diagram · No Style Control

AR · Vision understanding · Human

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

Rank #10 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,303
Percentile: 87.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: diagram. Source rank: #14. Votes: 2880. Organization: google. License: Proprietary.

87.1% percentile inside its fair comparison set

1,303Raw benchmark valueCI 1,291 - 1,314

Vision Arena · English · No Style Control

AR · Vision understanding · Human

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

Rank #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,293
Percentile: 92.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #11. Votes: 5347. Organization: google. License: Proprietary.

92.7% percentile inside its fair comparison set

1,293Raw benchmark valueCI 1,282 - 1,303

Vision Arena · Entity Recognition · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,358
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: entity_recognition. Source rank: #1. Votes: 243. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,358Raw benchmark valueCI 1,323 - 1,393

Vision Arena · Homework · No Style Control

AR · Vision understanding · Human

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

Rank #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,317
Percentile: 89.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: homework. Source rank: #11. Votes: 1795. Organization: google. License: Proprietary.

89.7% percentile inside its fair comparison set

1,317Raw benchmark valueCI 1,303 - 1,331

Vision Arena · Humor · No Style Control

AR · Vision understanding · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,339
Percentile: 98%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: humor. Source rank: #2. Votes: 564. Organization: google. License: Proprietary.

98% percentile inside its fair comparison set

1,339Raw benchmark valueCI 1,314 - 1,364

Vision Arena · Ocr · No Style Control

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,311
Percentile: 94.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: ocr. Source rank: #7. Votes: 8110. Organization: google. License: Proprietary.

94.3% percentile inside its fair comparison set

1,311Raw benchmark valueCI 1,303 - 1,320

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 #6 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 87.5%
Percentile: 91.4%
Last updated: recent
Eligibility: preview_model

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

91.4% percentile inside its fair comparison set

87.5%Raw benchmark valueCI 86% - 89.1%

VTB

SL · Vision understanding · Rubric

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

Rank #4

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 26.9%
Percentile: 72.7%
Last updated: recent
Eligibility: preview_model

72.7% percentile inside its fair comparison set

26.9%Raw benchmark value

MMMU-Pro

OFF · Vision understanding · Objective

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

Rank #3

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 81%
Percentile: 60%
Last updated: recent
Eligibility: preview_model

Verified against the current Gemini Pro public model page. Page did not expose a visible page date on April 29, 2026. Uses the published MMMU-Pro figure for Gemini 3 Pro Preview. Checked 2026-04-29. Verification: manual_public_page_verification.

60% percentile inside its fair comparison set

81%Raw benchmark value

Vision Arena · Creative Writing

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,373
Percentile: 100%
Last updated: archived
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #1. Votes: 766. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,373Raw benchmark valueCI 1,350 - 1,395

Vision Arena · Creative Writing · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,391
Percentile: 100%
Last updated: archived
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #1. Votes: 766. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,391Raw benchmark valueCI 1,368 - 1,413

Document understanding3 benchmarks62.2%

Document Arena

AR · Document understanding · Human

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

Rank #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,452
Percentile: 50%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #16. Votes: 10748. Organization: google. License: Proprietary.

50% percentile inside its fair comparison set

1,452Raw benchmark valueCI 1,443 - 1,461

Document Arena · No Style Control

AR · Document understanding · Human

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

Rank #15 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 41.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #18. Votes: 10748. Organization: google. License: Proprietary.

41.7% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,424 - 1,442

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 #4 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 69.1%
Percentile: 95%
Last updated: recent
Eligibility: preview_model

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

95% percentile inside its fair comparison set

69.1%Raw benchmark valueCI 67.3% - 70.9%

Safety1 benchmark7.7%

MASK

SL · Safety · Rubric

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

Rank #13

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 42.6%
Percentile: 7.7%
Last updated: recent
Eligibility: preview_model

7.7% percentile inside its fair comparison set

42.6%Raw benchmark value

Multilingual16 benchmarks98.3%

Text Arena · Chinese

AR · Multilingual · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,529
Percentile: 98.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #7. Votes: 2102. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,529Raw benchmark valueCI 1,515 - 1,542

Text Arena · French

AR · Multilingual · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,509
Percentile: 97.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: french. Source rank: #7. Votes: 890. Organization: google. License: Proprietary.

97.7% percentile inside its fair comparison set

1,509Raw benchmark valueCI 1,488 - 1,530

Text Arena · German

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: german. Source rank: #1. Votes: 764. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,520Raw benchmark valueCI 1,497 - 1,542

Text Arena · Japanese

AR · Multilingual · Human

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

Rank #3 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,507
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: japanese. Source rank: #3. Votes: 406. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,507Raw benchmark valueCI 1,475 - 1,540

Text Arena · Korean

AR · Multilingual · Human

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

Rank #10 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,452
Percentile: 95.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: korean. Source rank: #11. Votes: 667. Organization: google. License: Proprietary.

95.7% percentile inside its fair comparison set

1,452Raw benchmark valueCI 1,427 - 1,476

Text Arena · Russian

AR · Multilingual · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,503
Percentile: 98.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: russian. Source rank: #6. Votes: 4040. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,503Raw benchmark valueCI 1,493 - 1,512

Text Arena · Spanish

AR · Multilingual · Human

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

Rank #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,473
Percentile: 96.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: spanish. Source rank: #10. Votes: 1037. Organization: google. License: Proprietary.

96.7% percentile inside its fair comparison set

1,473Raw benchmark valueCI 1,454 - 1,493

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,529
Percentile: 98.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #9. Votes: 2102. Organization: google. License: Proprietary.

98.3% percentile inside its fair comparison set

1,529Raw benchmark valueCI 1,515 - 1,542

Text Arena · French · No Style Control

AR · Multilingual · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,497
Percentile: 97.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: french. Source rank: #8. Votes: 890. Organization: google. License: Proprietary.

97.7% percentile inside its fair comparison set

1,497Raw benchmark valueCI 1,476 - 1,518

Text Arena · German · No Style Control

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,515
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: german. Source rank: #1. Votes: 764. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,515Raw benchmark valueCI 1,492 - 1,538

Text Arena · Japanese · No Style Control

AR · Multilingual · Human

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

Rank #3 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,511
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: japanese. Source rank: #3. Votes: 406. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,511Raw benchmark valueCI 1,478 - 1,543

Text Arena · Korean · No Style Control

AR · Multilingual · Human

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

Rank #7 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,451
Percentile: 97.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: korean. Source rank: #7. Votes: 667. Organization: google. License: Proprietary.

97.1% percentile inside its fair comparison set

1,451Raw benchmark valueCI 1,427 - 1,476

Text Arena · Russian · No Style Control

AR · Multilingual · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 98.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: russian. Source rank: #6. Votes: 4040. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,486 - 1,505

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

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

Rank #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 96.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: spanish. Source rank: #10. Votes: 1037. Organization: google. License: Proprietary.

96.7% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,449 - 1,488

Vision Arena · Chinese

AR · Multilingual · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,351
Percentile: 98.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #2. Votes: 1400. Organization: google. License: Proprietary.

98.7% percentile inside its fair comparison set

1,351Raw benchmark valueCI 1,331 - 1,371

Vision Arena · Chinese · No Style Control

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,375
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #1. Votes: 1400. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,375Raw benchmark valueCI 1,355 - 1,395

Source links and registry checks

official

Google Gemini models docs

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Arena

Jun 20, 2026

source →

official

Artificial Analysis

Jun 20, 2026

source →

official

Terminal-Bench

Jun 20, 2026

source →

Model profile · Google

Gemini 3 Pro Preview

Closed weightsfrontier · registry tag 2026 preview

Thin verified coverage

Reads as thin verified coverage across the resolved source data.

Visible coverage: 34.2%
Verified coverage: 34.2%
Spread: n/a
Last verified: Jun 20, 2026

textcodevisiondocumentaudiosearch6 aliases48 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.

Thin verified coverageThis model currently reads as thin verified coverage across the resolved source data.

Chat / text26 benchmarks80.5%

Intelligence Index

AA · Chat / text · Combined

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

Rank #39 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 33
Percentile: 90.4%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `intelligenceIndex`.

90.4% percentile inside its fair comparison set

33Raw 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 #4 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 47.1%
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `omniscienceAccuracy`.

99% percentile inside its fair comparison set

47.1%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 #229 · Source label: Gemini 3 Pro Preview (high)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 9.1%
Percentile: 23.5%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `omniscienceNonHallucination`.

23.5% percentile inside its fair comparison set

9.1%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 #94 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 49.7%
Percentile: 70.5%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `ifbench`.

70.5% percentile inside its fair comparison set

49.7%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 #242 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $4.5 /1M tokens
Percentile: 13%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `price1mBlended0To3To1`.

13% percentile inside its fair comparison set

$4.5 /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 #237 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $2 /1M input tokens
Percentile: 17.8%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `price1mInputTokens`.

17.8% percentile inside its fair comparison set

$2 /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 #242 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: $12 /1M output tokens
Percentile: 13%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `price1mOutputTokens`.

13% percentile inside its fair comparison set

$12 /1M output tokensRaw benchmark value

Openness Index

AA · Chat / text · Combined

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

Rank #185 · Source label: Gemini 3 Pro Preview (high)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

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

7.5% percentile inside its fair comparison set

6Raw benchmark value

Text Arena

AR · Chat / text · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #8. Votes: 41314. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,486Raw benchmark valueCI 1,482 - 1,490

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,485
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #4. Votes: 6288. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,485Raw benchmark valueCI 1,476 - 1,493

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 #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,489
Percentile: 98.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #8. Votes: 18965. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,489Raw benchmark valueCI 1,484 - 1,494

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,501
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: exclude_ties. Source rank: #6. Votes: 29671. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,501Raw benchmark valueCI 1,496 - 1,507

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 #7 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,504
Percentile: 98.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts. Source rank: #9. Votes: 22462. Organization: google. License: Proprietary.

98.2% percentile inside its fair comparison set

1,504Raw benchmark valueCI 1,499 - 1,509

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,502
Percentile: 96.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts_english. Source rank: #14. Votes: 10730. Organization: google. License: Proprietary.

96.6% percentile inside its fair comparison set

1,502Raw benchmark valueCI 1,496 - 1,509

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 #11 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,474
Percentile: 96.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: instruction_following. Source rank: #14. Votes: 11193. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,474Raw benchmark valueCI 1,468 - 1,481

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 #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,492
Percentile: 97.4%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: longer_query. Source rank: #13. Votes: 10490. Organization: google. License: Proprietary.

97.4% percentile inside its fair comparison set

1,492Raw benchmark valueCI 1,486 - 1,499

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 #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 98.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: multi_turn. Source rank: #8. Votes: 6769. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,487 - 1,503

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 #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 98.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #8. Votes: 41314. Organization: google. License: Proprietary.

98.5% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,476 - 1,483

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #5. Votes: 6288. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,474 - 1,490

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,479
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #6. Votes: 18965. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,479Raw benchmark valueCI 1,474 - 1,485

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,491
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: exclude_ties. Source rank: #5. Votes: 29671. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,491Raw benchmark valueCI 1,486 - 1,496

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 #11 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 96.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts. Source rank: #14. Votes: 22462. Organization: google. License: Proprietary.

96.9% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,477 - 1,487

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 #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 95.4%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: hard_prompts_english. Source rank: #19. Votes: 10730. Organization: google. License: Proprietary.

95.4% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,472 - 1,484

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 #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,459
Percentile: 95.4%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: instruction_following. Source rank: #21. Votes: 11193. Organization: google. License: Proprietary.

95.4% percentile inside its fair comparison set

1,459Raw benchmark valueCI 1,453 - 1,465

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 #17 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,473
Percentile: 94.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: longer_query. Source rank: #22. Votes: 10490. Organization: google. License: Proprietary.

94.7% percentile inside its fair comparison set

1,473Raw benchmark valueCI 1,467 - 1,480

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 #7 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,483
Percentile: 98.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: multi_turn. Source rank: #9. Votes: 6769. Organization: google. License: Proprietary.

98.1% percentile inside its fair comparison set

1,483Raw benchmark valueCI 1,475 - 1,490

Coding16 benchmarks67.7%

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 #46 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 34.1%
Percentile: 85.1%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `terminalbenchHard`.

85.1% percentile inside its fair comparison set

34.1%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 #9 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 49.9%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `scicode`.

97.8% percentile inside its fair comparison set

49.9%Raw 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 #23 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,439
Percentile: 69.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #29. Votes: 17158. Organization: google. License: Proprietary.

69.9% percentile inside its fair comparison set

1,439Raw benchmark valueCI 1,431 - 1,446

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 #23 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,439
Percentile: 69.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: webdev. Source rank: #29. Votes: 17158. Organization: google. License: Proprietary.

69.9% percentile inside its fair comparison set

1,439Raw benchmark valueCI 1,431 - 1,446

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 #20 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 74%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: webdev-html. Source rank: #26. Votes: 13755. Organization: google. License: Proprietary.

74% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,453 - 1,472

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 #28 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,402
Percentile: 54.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: webdev-react. Source rank: #35. Votes: 3394. Organization: google. License: Proprietary.

54.2% percentile inside its fair comparison set

1,402Raw benchmark valueCI 1,391 - 1,413

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 #12 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 86.4%
Percentile: 87.8%
Last updated: recent
Eligibility: preview_model

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

87.8% percentile inside its fair comparison set

86.4%Raw benchmark valueCI 84.5% - 88.3%

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 #15 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 76.4%
Percentile: 75.9%
Last updated: recent
Eligibility: preview_model

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

75.9% percentile inside its fair comparison set

76.4%Raw benchmark valueCI 72.7% - 80.1%

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 #36 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 14.3%
Percentile: 28.6%
Last updated: recent
Eligibility: preview_model

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

28.6% percentile inside its fair comparison set

14.3%Raw benchmark valueCI 8.3% - 20.3%

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 #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,519
Percentile: 95.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: coding. Source rank: #20. Votes: 8573. Organization: google. License: Proprietary.

95.3% percentile inside its fair comparison set

1,519Raw benchmark valueCI 1,512 - 1,526

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 #20 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,483
Percentile: 94.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: coding. Source rank: #25. Votes: 8573. Organization: google. License: Proprietary.

94.1% percentile inside its fair comparison set

1,483Raw benchmark valueCI 1,476 - 1,490

IOI

VALS-AI · Coding · Objective

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

Rank #8 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 38.8%
Percentile: 84.1%
Last updated: recent
Eligibility: preview_model

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

84.1% percentile inside its fair comparison set

38.8%Raw benchmark valueCI 16.3% - 61.4%

Code Arena · Image To Webdev

AR · Coding · Human

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

Rank #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,453
Percentile: 52.9%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: image_to_webdev. Source rank: #12. Votes: 1091. Organization: google. License: Proprietary.

52.9% percentile inside its fair comparison set

1,453Raw benchmark valueCI 1,433 - 1,473

Terminal-Bench 2.0

OFF · Coding · Objective

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

Rank #7

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 56.9%
Percentile: 0%
Last updated: recent
Eligibility: preview_model

0% percentile inside its fair comparison set

56.9%Raw benchmark value

SWE-Bench Verified

OFF · Coding · Objective

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

Rank #5

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 76.2%
Percentile: 20%
Last updated: recent
Eligibility: preview_model

20% percentile inside its fair comparison set

76.2%Raw benchmark value

Terminal-Bench 2.0

TERMINAL-BENCH · Coding · Objective

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

Rank #3

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Terminal-Bench
Raw value: 69.4%
Percentile: 93.3%
Last updated: archived
Eligibility: preview_model

93.3% percentile inside its fair comparison set

69.4%Raw benchmark value

Reasoning / math / science11 benchmarks77%

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 #18 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 27.6%
Percentile: 95.4%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `hle`.

95.4% percentile inside its fair comparison set

27.6%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 #12 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 88.7%
Percentile: 97.1%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `gpqa`.

97.1% percentile inside its fair comparison set

88.7%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 #108 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

Parsed from Artificial Analysis public leaderboard field `critpt`.

65.2% percentile inside its fair comparison set

0%Raw benchmark value

ProofBench

VALS-AI · Reasoning / math / science · Objective

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

Rank #16 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 20%
Percentile: 57.1%
Last updated: recent
Eligibility: preview_model

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

57.1% percentile inside its fair comparison set

20%Raw benchmark valueCI 12.2% - 27.8%

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 #7 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 91.7%
Percentile: 93.3%
Last updated: recent
Eligibility: preview_model

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

93.3% percentile inside its fair comparison set

91.7%Raw benchmark valueCI 88.9% - 94.4%

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 #3 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 90.1%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

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

97.8% percentile inside its fair comparison set

90.1%Raw benchmark valueCI 89.5% - 90.7%

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 #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 96.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: math. Source rank: #17. Votes: 2652. Organization: google. License: Proprietary.

96.2% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,466 - 1,489

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,476
Percentile: 96.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: math. Source rank: #15. Votes: 2652. Organization: google. License: Proprietary.

96.5% percentile inside its fair comparison set

1,476Raw benchmark valueCI 1,465 - 1,487

TutorBench

SL · Reasoning / math / science · Rubric

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

Rank #7

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 53.7%
Percentile: 40%
Last updated: recent
Eligibility: preview_model

40% percentile inside its fair comparison set

53.7%Raw benchmark value

MultiNRC

SL · Reasoning / math / science · Rubric

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

Rank #3

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 59%
Percentile: 80%
Last updated: recent
Eligibility: preview_model

80% percentile inside its fair comparison set

59%Raw benchmark value

Humanity's Last Exam

OFF · Reasoning / math / science · Objective

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

Rank #6

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 37.5%
Percentile: 28.6%
Last updated: recent
Eligibility: preview_model

28.6% percentile inside its fair comparison set

37.5%Raw benchmark value

Professional reasoning25 benchmarks85.7%

LegalBench

VALS-AI · Professional reasoning · Objective

Academic legal reasoning tasks.

Rank #3 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 87%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

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

97.8% percentile inside its fair comparison set

87%Raw benchmark valueCI 86.3% - 87.7%

TaxEval v2

VALS-AI · Professional reasoning · Objective

Answer quality on tax questions and responses.

Rank #42 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 72.6%
Percentile: 54.9%
Last updated: recent
Eligibility: preview_model

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

54.9% percentile inside its fair comparison set

72.6%Raw benchmark valueCI 70.9% - 74.3%

MedCode

VALS-AI · Professional reasoning · Objective

Medical billing support and coding tasks.

Rank #8 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 52.2%
Percentile: 86.3%
Last updated: recent
Eligibility: preview_model

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

86.3% percentile inside its fair comparison set

52.2%Raw benchmark valueCI 48.1% - 56.3%

MedScribe

VALS-AI · Professional reasoning · Objective

Administrative documentation support for doctors.

Rank #43 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 72%
Percentile: 16%
Last updated: recent
Eligibility: preview_model

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

16% percentile inside its fair comparison set

72%Raw benchmark valueCI 68.3% - 75.8%

Text Arena · Expert

AR · Professional reasoning · Human

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

Rank #14 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,503
Percentile: 95.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: expert. Source rank: #19. Votes: 2532. Organization: google. License: Proprietary.

95.3% percentile inside its fair comparison set

1,503Raw benchmark valueCI 1,491 - 1,515

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,475
Percentile: 96.5%
Last updated: recent
Eligibility: preview_model

96.5% percentile inside its fair comparison set

1,475Raw benchmark valueCI 1,468 - 1,483

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,468
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_entertainment_and_sports_and_media. Source rank: #6. Votes: 7675. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,468Raw benchmark valueCI 1,461 - 1,476

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,502
Percentile: 98.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_legal_and_government. Source rank: #7. Votes: 2904. Organization: google. License: Proprietary.

98.3% percentile inside its fair comparison set

1,502Raw benchmark valueCI 1,491 - 1,513

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 #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,499
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_life_and_physical_and_social_science. Source rank: #10. Votes: 6719. Organization: google. License: Proprietary.

97.8% percentile inside its fair comparison set

1,499Raw benchmark valueCI 1,491 - 1,506

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

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

Rank #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,482
Percentile: 95.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_mathematical. Source rank: #19. Votes: 1941. Organization: google. License: Proprietary.

95.1% percentile inside its fair comparison set

1,482Raw benchmark valueCI 1,468 - 1,495

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

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

Rank #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,508
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_medicine_and_healthcare. Source rank: #6. Votes: 2532. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,508Raw benchmark valueCI 1,496 - 1,521

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 #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,511
Percentile: 96.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_software_and_it_services. Source rank: #15. Votes: 14254. Organization: google. License: Proprietary.

96.6% percentile inside its fair comparison set

1,511Raw benchmark valueCI 1,505 - 1,517

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,481
Percentile: 99.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_writing_and_literature_and_language. Source rank: #6. Votes: 9248. Organization: google. License: Proprietary.

99.1% percentile inside its fair comparison set

1,481Raw benchmark valueCI 1,474 - 1,488

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

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

Rank #18 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,480
Percentile: 93.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: expert. Source rank: #24. Votes: 2532. Organization: google. License: Proprietary.

93.8% percentile inside its fair comparison set

1,480Raw benchmark valueCI 1,468 - 1,492

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 #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,460
Percentile: 96.2%
Last updated: recent
Eligibility: preview_model

96.2% percentile inside its fair comparison set

1,460Raw benchmark valueCI 1,453 - 1,468

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 #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,464
Percentile: 98.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_entertainment_and_sports_and_media. Source rank: #7. Votes: 7675. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,464Raw benchmark valueCI 1,456 - 1,471

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,493
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_legal_and_government. Source rank: #5. Votes: 2904. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,493Raw benchmark valueCI 1,482 - 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 #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,486
Percentile: 97.5%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_life_and_physical_and_social_science. Source rank: #11. Votes: 6719. Organization: google. License: Proprietary.

97.5% percentile inside its fair comparison set

1,486Raw benchmark valueCI 1,478 - 1,494

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

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

Rank #16 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,475
Percentile: 95.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_mathematical. Source rank: #19. Votes: 1941. Organization: google. License: Proprietary.

95.1% percentile inside its fair comparison set

1,475Raw benchmark valueCI 1,462 - 1,488

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 #4 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,492
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_medicine_and_healthcare. Source rank: #5. Votes: 2532. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,492Raw benchmark valueCI 1,480 - 1,504

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 #15 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 95.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_software_and_it_services. Source rank: #18. Votes: 14254. Organization: google. License: Proprietary.

95.7% percentile inside its fair comparison set

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

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 #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,478
Percentile: 98.8%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: industry_writing_and_literature_and_language. Source rank: #7. Votes: 9248. Organization: google. License: Proprietary.

98.8% percentile inside its fair comparison set

1,478Raw benchmark valueCI 1,471 - 1,485

SAGE

VALS-AI · Professional reasoning · Objective

Student Assessment with Generative Evaluation.

Rank #17 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 47.6%
Percentile: 64.4%
Last updated: recent
Eligibility: preview_model

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

64.4% percentile inside its fair comparison set

47.6%Raw benchmark valueCI 41% - 54.2%

PRBench Legal

SL · Professional reasoning · Rubric

Applied legal reasoning on professional-domain tasks.

Rank #11

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 40.6%
Percentile: 16.7%
Last updated: recent
Eligibility: preview_model

16.7% percentile inside its fair comparison set

40.6%Raw benchmark value

Poker Agent

VALS-AI · Professional reasoning · Objective

Agent profit in poker-style strategic play.

Rank #8 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 1,078.9 score
Percentile: 56.3%
Last updated: archived
Eligibility: preview_model

Parsed from Vals AI BenchmarkView overall scores. Vals slug: poker_agent; provider: unknown.

56.3% percentile inside its fair comparison set

1,078.9 scoreRaw benchmark valueCI 1,078.9 score - 1,078.9 score

Search / tool use4 benchmarks56.2%

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 #100 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 68.1%
Percentile: 68%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `tau2`.

68% percentile inside its fair comparison set

68.1%Raw benchmark value

Search Arena

AR · Search / tool use · Human

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

Rank #7 · Source label: gemini-3-pro-grounding

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,201
Percentile: 80%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro-grounding`. Category: overall. Source rank: #7. Votes: 37327. Organization: google. License: Proprietary.

80% percentile inside its fair comparison set

1,201Raw benchmark valueCI 1,195 - 1,206

Search Arena · No Style Control

AR · Search / tool use · Human

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

Rank #8 · Source label: gemini-3-pro-grounding

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,207
Percentile: 76.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro-grounding`. Category: overall. Source rank: #8. Votes: 37327. Organization: google. License: Proprietary.

76.7% percentile inside its fair comparison set

1,207Raw benchmark valueCI 1,201 - 1,212

BrowseComp

OFF · Search / tool use · Objective

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

Rank #7

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 59.2%
Percentile: 0%
Last updated: recent
Eligibility: preview_model

0% percentile inside its fair comparison set

59.2%Raw benchmark value

Long context3 benchmarks54.5%

Long Context Reasoning

AA · Long context · Objective

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

Rank #16 · Source label: Gemini 3 Pro Preview (low)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 67.3%
Percentile: 95.2%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `lcr`.

95.2% percentile inside its fair comparison set

67.3%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 #30 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 63.7%
Percentile: 68.2%
Last updated: recent
Eligibility: preview_model

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

68.2% percentile inside its fair comparison set

63.7%Raw benchmark valueCI 61.8% - 65.5%

MRCR v2

OFF · Long context · Objective

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

Rank #2

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 77%
Percentile: 0%
Last updated: recent
Eligibility: preview_model

0% percentile inside its fair comparison set

77%Raw benchmark value

Vision understanding24 benchmarks92.8%

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 #4 · Source label: Gemini 3 Pro Preview (high)

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Artificial Analysis
Raw value: 80.2%
Percentile: 97.8%
Last updated: recent
Eligibility: preview_model

Parsed from Artificial Analysis public leaderboard field `mmmuPro`.

97.8% percentile inside its fair comparison set

80.2%Raw benchmark value

Vision Arena

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,290
Percentile: 96.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #7. Votes: 13200. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,290Raw benchmark valueCI 1,282 - 1,297

Vision Arena · Captioning

AR · Vision understanding · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,273
Percentile: 96.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: captioning. Source rank: #2. Votes: 222. Organization: google. License: Proprietary.

96.2% percentile inside its fair comparison set

1,273Raw benchmark valueCI 1,238 - 1,309

Vision Arena · Creative Writing Vision

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,333
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing_vision. Source rank: #1. Votes: 1039. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,333Raw benchmark valueCI 1,314 - 1,352

Vision Arena · Diagram

AR · Vision understanding · Human

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

Rank #12 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,301
Percentile: 84.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: diagram. Source rank: #16. Votes: 2880. Organization: google. License: Proprietary.

84.3% percentile inside its fair comparison set

1,301Raw benchmark valueCI 1,290 - 1,313

Vision Arena · English

AR · Vision understanding · Human

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

Rank #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,276
Percentile: 89%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #15. Votes: 5347. Organization: google. License: Proprietary.

89% percentile inside its fair comparison set

1,276Raw benchmark valueCI 1,265 - 1,286

Vision Arena · Entity Recognition

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,305
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: entity_recognition. Source rank: #1. Votes: 243. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,305Raw benchmark valueCI 1,269 - 1,340

Vision Arena · Homework

AR · Vision understanding · Human

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

Rank #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,311
Percentile: 88.2%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: homework. Source rank: #12. Votes: 1795. Organization: google. License: Proprietary.

88.2% percentile inside its fair comparison set

1,311Raw benchmark valueCI 1,296 - 1,325

Vision Arena · Humor

AR · Vision understanding · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,316
Percentile: 98%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: humor. Source rank: #2. Votes: 564. Organization: google. License: Proprietary.

98% percentile inside its fair comparison set

1,316Raw benchmark valueCI 1,291 - 1,341

Vision Arena · Ocr

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,303
Percentile: 94.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: ocr. Source rank: #7. Votes: 8110. Organization: google. License: Proprietary.

94.3% percentile inside its fair comparison set

1,303Raw benchmark valueCI 1,295 - 1,312

Vision Arena · No Style Control

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,305
Percentile: 96.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #7. Votes: 13200. Organization: google. License: Proprietary.

96.3% percentile inside its fair comparison set

1,305Raw benchmark valueCI 1,298 - 1,313

Vision Arena · Captioning · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,321
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: captioning. Source rank: #1. Votes: 222. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,321Raw benchmark valueCI 1,287 - 1,355

Vision Arena · Creative Writing Vision · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,350
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing_vision. Source rank: #1. Votes: 1039. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,350Raw benchmark valueCI 1,331 - 1,369

Vision Arena · Diagram · No Style Control

AR · Vision understanding · Human

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

Rank #10 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,303
Percentile: 87.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: diagram. Source rank: #14. Votes: 2880. Organization: google. License: Proprietary.

87.1% percentile inside its fair comparison set

1,303Raw benchmark valueCI 1,291 - 1,314

Vision Arena · English · No Style Control

AR · Vision understanding · Human

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

Rank #9 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,293
Percentile: 92.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: english. Source rank: #11. Votes: 5347. Organization: google. License: Proprietary.

92.7% percentile inside its fair comparison set

1,293Raw benchmark valueCI 1,282 - 1,303

Vision Arena · Entity Recognition · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,358
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: entity_recognition. Source rank: #1. Votes: 243. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,358Raw benchmark valueCI 1,323 - 1,393

Vision Arena · Homework · No Style Control

AR · Vision understanding · Human

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

Rank #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,317
Percentile: 89.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: homework. Source rank: #11. Votes: 1795. Organization: google. License: Proprietary.

89.7% percentile inside its fair comparison set

1,317Raw benchmark valueCI 1,303 - 1,331

Vision Arena · Humor · No Style Control

AR · Vision understanding · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,339
Percentile: 98%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: humor. Source rank: #2. Votes: 564. Organization: google. License: Proprietary.

98% percentile inside its fair comparison set

1,339Raw benchmark valueCI 1,314 - 1,364

Vision Arena · Ocr · No Style Control

AR · Vision understanding · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,311
Percentile: 94.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: ocr. Source rank: #7. Votes: 8110. Organization: google. License: Proprietary.

94.3% percentile inside its fair comparison set

1,311Raw benchmark valueCI 1,303 - 1,320

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 #6 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 87.5%
Percentile: 91.4%
Last updated: recent
Eligibility: preview_model

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

91.4% percentile inside its fair comparison set

87.5%Raw benchmark valueCI 86% - 89.1%

VTB

SL · Vision understanding · Rubric

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

Rank #4

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 26.9%
Percentile: 72.7%
Last updated: recent
Eligibility: preview_model

72.7% percentile inside its fair comparison set

26.9%Raw benchmark value

MMMU-Pro

OFF · Vision understanding · Objective

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

Rank #3

Official company resultmanual verifiedmanual verifiedBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Provider official evals
Raw value: 81%
Percentile: 60%
Last updated: recent
Eligibility: preview_model

60% percentile inside its fair comparison set

81%Raw benchmark value

Vision Arena · Creative Writing

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,373
Percentile: 100%
Last updated: archived
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #1. Votes: 766. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,373Raw benchmark valueCI 1,350 - 1,395

Vision Arena · Creative Writing · No Style Control

AR · Vision understanding · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,391
Percentile: 100%
Last updated: archived
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: creative_writing. Source rank: #1. Votes: 766. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,391Raw benchmark valueCI 1,368 - 1,413

Document understanding3 benchmarks62.2%

Document Arena

AR · Document understanding · Human

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

Rank #13 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,452
Percentile: 50%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #16. Votes: 10748. Organization: google. License: Proprietary.

50% percentile inside its fair comparison set

1,452Raw benchmark valueCI 1,443 - 1,461

Document Arena · No Style Control

AR · Document understanding · Human

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

Rank #15 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 41.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: overall. Source rank: #18. Votes: 10748. Organization: google. License: Proprietary.

41.7% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,424 - 1,442

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 #4 · Source label: google/gemini-3-pro-preview

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Vals AI
Raw value: 69.1%
Percentile: 95%
Last updated: recent
Eligibility: preview_model

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

95% percentile inside its fair comparison set

69.1%Raw benchmark valueCI 67.3% - 70.9%

Safety1 benchmark7.7%

MASK

SL · Safety · Rubric

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

Rank #13

verified runtimeexact directBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Scale Labs
Raw value: 42.6%
Percentile: 7.7%
Last updated: recent
Eligibility: preview_model

7.7% percentile inside its fair comparison set

42.6%Raw benchmark value

Multilingual16 benchmarks98.3%

Text Arena · Chinese

AR · Multilingual · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,529
Percentile: 98.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #7. Votes: 2102. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,529Raw benchmark valueCI 1,515 - 1,542

Text Arena · French

AR · Multilingual · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,509
Percentile: 97.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: french. Source rank: #7. Votes: 890. Organization: google. License: Proprietary.

97.7% percentile inside its fair comparison set

1,509Raw benchmark valueCI 1,488 - 1,530

Text Arena · German

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

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

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: german. Source rank: #1. Votes: 764. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,520Raw benchmark valueCI 1,497 - 1,542

Text Arena · Japanese

AR · Multilingual · Human

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

Rank #3 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,507
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: japanese. Source rank: #3. Votes: 406. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,507Raw benchmark valueCI 1,475 - 1,540

Text Arena · Korean

AR · Multilingual · Human

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

Rank #10 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,452
Percentile: 95.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: korean. Source rank: #11. Votes: 667. Organization: google. License: Proprietary.

95.7% percentile inside its fair comparison set

1,452Raw benchmark valueCI 1,427 - 1,476

Text Arena · Russian

AR · Multilingual · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,503
Percentile: 98.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: russian. Source rank: #6. Votes: 4040. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,503Raw benchmark valueCI 1,493 - 1,512

Text Arena · Spanish

AR · Multilingual · Human

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

Rank #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,473
Percentile: 96.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: spanish. Source rank: #10. Votes: 1037. Organization: google. License: Proprietary.

96.7% percentile inside its fair comparison set

1,473Raw benchmark valueCI 1,454 - 1,493

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,529
Percentile: 98.3%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #9. Votes: 2102. Organization: google. License: Proprietary.

98.3% percentile inside its fair comparison set

1,529Raw benchmark valueCI 1,515 - 1,542

Text Arena · French · No Style Control

AR · Multilingual · Human

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

Rank #6 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,497
Percentile: 97.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: french. Source rank: #8. Votes: 890. Organization: google. License: Proprietary.

97.7% percentile inside its fair comparison set

1,497Raw benchmark valueCI 1,476 - 1,518

Text Arena · German · No Style Control

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,515
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: german. Source rank: #1. Votes: 764. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,515Raw benchmark valueCI 1,492 - 1,538

Text Arena · Japanese · No Style Control

AR · Multilingual · Human

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

Rank #3 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,511
Percentile: 99%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: japanese. Source rank: #3. Votes: 406. Organization: google. License: Proprietary.

99% percentile inside its fair comparison set

1,511Raw benchmark valueCI 1,478 - 1,543

Text Arena · Korean · No Style Control

AR · Multilingual · Human

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

Rank #7 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,451
Percentile: 97.1%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: korean. Source rank: #7. Votes: 667. Organization: google. License: Proprietary.

97.1% percentile inside its fair comparison set

1,451Raw benchmark valueCI 1,427 - 1,476

Text Arena · Russian · No Style Control

AR · Multilingual · Human

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

Rank #5 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,495
Percentile: 98.6%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: russian. Source rank: #6. Votes: 4040. Organization: google. License: Proprietary.

98.6% percentile inside its fair comparison set

1,495Raw benchmark valueCI 1,486 - 1,505

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

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

Rank #8 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,469
Percentile: 96.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: spanish. Source rank: #10. Votes: 1037. Organization: google. License: Proprietary.

96.7% percentile inside its fair comparison set

1,469Raw benchmark valueCI 1,449 - 1,488

Vision Arena · Chinese

AR · Multilingual · Human

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

Rank #2 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,351
Percentile: 98.7%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #2. Votes: 1400. Organization: google. License: Proprietary.

98.7% percentile inside its fair comparison set

1,351Raw benchmark valueCI 1,331 - 1,371

Vision Arena · Chinese · No Style Control

AR · Multilingual · Human

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

Rank #1 · Source label: gemini-3-pro

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,375
Percentile: 100%
Last updated: recent
Eligibility: preview_model

Parsed from Arena leaderboard dataset row `gemini-3-pro`. Category: chinese. Source rank: #1. Votes: 1400. Organization: google. License: Proprietary.

100% percentile inside its fair comparison set

1,375Raw benchmark valueCI 1,355 - 1,395