Skip to main content

Model profile

Gemini 3 Pro

GoogleEstablishedReleased Nov 18, 2025
Data verified
Overall Score
67.73Public #19 of 200Verified #17 of 99
Arena Elo
1486
Eligible category ranks
4of 8
Price (1M tokens)
$2 in / $12 out
API pricing
Speed
109tok/s
Context
2M

Evidence coverage

27 of 321 tracked benchmarks are published. 9 are verified and 18 provisional. 8 of 8 categories are measured.

Updated July 20, 2026Methodology
Published / tracked
27 / 321
Verified
9
Provisional
18
Categories with evidence
8 / 8

Evidence by category

  • Agentic3 benchmarks
    Mixed evidence
  • Coding3 benchmarks
    Mixed evidence
  • Reasoning3 benchmarks
    Mixed evidence
  • Knowledge7 benchmarks
    Reported
  • Math2 benchmarks
    Verified
  • Multilingual1 benchmark
    Reported
  • Multimodal7 benchmarks
    Mixed evidence
  • Inst. Following1 benchmark
    Reported
ProprietaryNon-Reasoning
Confidence:
Medium
base

Gemini 3 Pro ranks #19 out of 200 models on the public leaderboard with an overall score of 67.73/100. It also ranks #17 out of 99 on the verified leaderboard. This places it in the mid-tier of AI models, with strengths in specific benchmark categories.

Gemini 3 Pro is a proprietary model with a 2M token context window. It processes queries without explicit chain-of-thought reasoning, offering faster response times and lower token usage.

Gemini 3 Pro sits inside the Gemini 3 Pro family alongside Gemini 3 Pro Deep Think. This profile currently has 27 of 321 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Multimodal & Grounded (#9), while its weakest is Knowledge (#27). This performance profile makes it particularly strong for screenshots, documents, charts, and grounded multimodal workflows.

Peer position

Exact provisional scores and ranks for the closest listed peers. A score can appear before a model clears the evidence threshold for a rank, so equal scores can have different rank states.

Range 67.0167.78

  1. MiMo-V2-Pro
    Xiaomi
    #1767.78
    MiMo-V2-Pro is #17 with a score of 67.78.
    Compare
  2. GLM-5.1
    Z.AI
    #1867.74
    GLM-5.1 is #18 with a score of 67.74.
    Compare
  3. Gemini 3 ProCurrent model
    Google
    #1967.73
    Gemini 3 Pro is #19 with a score of 67.73.
  4. Inkling
    Thinking Machines Lab
    #2067.54
    Inkling is #20 with a score of 67.54.
    Compare
  5. Qwen3.7 Plus
    Alibaba
    #2167.22
    Qwen3.7 Plus is #21 with a score of 67.22.
    Compare
  6. GPT-5.6 Luna
    OpenAI
    #2267.17
    GPT-5.6 Luna is #22 with a score of 67.17.
    Compare
  7. GPT-5.2 Pro
    OpenAI
    #2367.01
    GPT-5.2 Pro is #23 with a score of 67.01.
    Compare

Category percentile

More

Relative position among models eligible for each sourced category. A higher percentile means a stronger position within that category's ranked cohort; 100 is highest.

  1. Multimodal71%
    Eligible cohort rank #9 of 29Category score 73.2
  2. Agentic90%
    Eligible cohort rank #13 of 119Category score 59.7
  3. Coding88%
    Eligible cohort rank #16 of 122Category score 62.8
  4. Knowledge49%
    Eligible cohort rank #27 of 52Category score 71.3

Category evidence

Scores and ranks appear only where this model has published benchmark evidence. Categories without displayable source records remain not measured.

Category scores, ranks, weighting, benchmark coverage, and evidence status
CategoryScore
AgenticRank #13 of 119Percentile 90thWeight 22%3 benchmarksMixed sources59.7
CodingRank #16 of 122Percentile 88thWeight 20%3 benchmarksMixed sources62.8
ReasoningRank Not rankedWeight 17%3 benchmarksMixed sources45.4
KnowledgeRank #27 of 52Percentile 49thWeight 12%7 benchmarksReported71.3
MathRank Not rankedWeight 5%2 benchmarksVerified56.3
MultilingualRank Not rankedWeight 7%1 benchmarkReported85.0
MultimodalRank #9 of 29Percentile 71stWeight 12%7 benchmarksMixed sources73.2
Inst. FollowingRank Not rankedWeight 5%1 benchmarkReported80.7

Chatbot Arena performance

Scroll horizontally to inspect confidence intervals and vote counts.

Chatbot Arena Elo, confidence interval, and vote count by evaluation view
ViewEloConfidence intervalVotes
Text Overall1486±3.941,290
Coding1519±7.18,574
Math1478±11.42,645
Instruction Following1474±6.411,179
Creative Writing1485±8.36,285
Multi-turn1495±7.96,766
Hard Prompts1504±5.022,449
Hard Prompts (English)1502±6.510,718
Longer Query1492±6.610,481

Benchmark Details

Rows below have a displayable published verification record. Each source link and provenance note remains in the page HTML while its category is closed. Source-unverified manual rows and generated rows stay hidden.

Agentic3 benchmarks
τ²-bench resultsReported

τ²-Bench Tool-Agent-User Evaluation

87.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Gert LabsBenchmark exact

Gert Labs Composite Game Benchmark

63.23%Display only
Source: Gert Labs rankingsProvenance: Gert Labs reports this composite leaderboard score in the public rankings API. BenchLM scales the source gscore from 0-1 to 0-100 and stores it as a display-only agentic benchmark.
JobBenchBenchmark exact
11.4%Display only
Source: JobBench paperProvenance: JobBench reports Gemini-3-Pro under OpenCode on the main-set leaderboard. BenchLM stores the reported main-set score.
Coding3 benchmarks
Vibe Code BenchBenchmark exact

Vibe Code Bench v1.1

14.30%Display only
Source: Vals AI: Vibe Code Bench v1.1Provenance: Vals Vibe Code Bench v1.1 reports this exact row under google/gemini-3-pro-preview; BenchLM stores it on the local vibeCodeBench key.
AA-SciCodeReported

Artificial Analysis SciCode

56.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA LiveCodeBenchReported

Artificial Analysis LiveCodeBench

91.7%Display only
Source: Artificial Analysis: livecodebench leaderboardProvenance: Display-only row synced from the current Artificial Analysis evaluation leaderboard. It is excluded from BenchLM weighted scoring.
Reasoning3 benchmarks
ARC-AGI-2Provider exact

Abstraction and Reasoning Corpus for AGI v2

31.1%Weighted 31%
Source: Google DeepMind: Gemini ProProvenance: Provider exact
AA-LCRReported

Artificial Analysis Long Context Reasoning

70.7%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
CritPtReported

Critical Physics Tasks

9.1%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Knowledge7 benchmarks
Artificial Analysis Intelligence IndexReported
39.5%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-GPQA DiamondReported

Artificial Analysis GPQA Diamond

90.8%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-HLEReported

Artificial Analysis Humanity's Last Exam

37.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience IndexReported

Artificial Analysis Omniscience Index

15.8%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience AccuracyReported

Artificial Analysis Omniscience Accuracy

55.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA-Omniscience Hallucination RateReported

Artificial Analysis Omniscience Hallucination Rate

90.9%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
AA MMLU-ProReported

Artificial Analysis MMLU-Pro

89.8%Display only
Source: Artificial Analysis: mmlu-pro leaderboardProvenance: Display-only row synced from the current Artificial Analysis evaluation leaderboard. It is excluded from BenchLM weighted scoring.
Math2 benchmarks
FrontierMath v2 (Tiers 1-3)Benchmark exact

FrontierMath v2 Tiers 1-3

37.600%Weighted 30%
Source: Epoch AI FrontierMath v2 leaderboardProvenance: Epoch AI reports FrontierMath v2 Tiers 1-3 at 37.6% for gemini-3-pro-preview. BenchLM selects the highest published thinking effort for the model and stores the v2 benchmark slice separately.
FrontierMath v2 (Tier 4)Benchmark exact

FrontierMath v2 Tier 4

18.750%Weighted 10%
Source: Epoch AI FrontierMath v2 leaderboardProvenance: Epoch AI reports FrontierMath v2 Tier 4 at 18.75% for gemini-3-pro-preview. BenchLM selects the highest published thinking effort for the model and stores the v2 benchmark slice separately.
Multilingual1 benchmark
AA Global-MMLU-LiteReported

Artificial Analysis Global-MMLU-Lite

92.2%Display only
Source: Artificial Analysis: global-mmlu-lite leaderboardProvenance: Display-only row synced from the current Artificial Analysis evaluation leaderboard. It is excluded from BenchLM weighted scoring.
Multimodal7 benchmarks
MMMU-ProProvider exact

Massive Multi-discipline Multimodal Understanding Pro

81%Weighted 45%
Source: Google: Gemini 3Provenance: Provider exact
CharXivSecondary exact

CharXiv Reasoning

81.4%Weighted 25%
Source: Qwen3.6-Plus multimodal comparison tableProvenance: Secondary exact
MathVisionSecondary exact
86.6%Display only
Source: Qwen3.6-Plus multimodal comparison tableProvenance: Secondary exact
VideoMMMUProvider exact
87.6%Display only
Source: Google: Gemini 3Provenance: Provider exact
ScreenSpot ProProvider exact
72.7%Display only
Source: Google DeepMind: Gemini FlashProvenance: Provider exact
V*Secondary exact
88.0%Display only
Source: Qwen3.6-Plus multimodal comparison tableProvenance: Secondary exact
AA-MMMU-ProReported

Artificial Analysis MMMU-Pro

80.2%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.
Inst. Following1 benchmark
AA-IFBenchReported

Artificial Analysis IFBench

70.4%Display only
Source: Artificial Analysis model benchmarksProvenance: Display-only row synced from the current Artificial Analysis model payload. It is excluded from BenchLM weighted scoring.

Gemini 3 Pro Family

Base entry

Frequently Asked Questions

How does Gemini 3 Pro perform overall in AI benchmarks?

Gemini 3 Pro currently ranks #19 out of 200 models on BenchLM's provisional leaderboard with an overall score of 67.73. It also ranks #17 out of 99 on the verified leaderboard. It is created by Google. Its published context window is 2M.

Is Gemini 3 Pro good for knowledge and understanding?

Gemini 3 Pro ranks #27 out of 52 models in knowledge and understanding benchmarks with an average score of 71.3. There are stronger options in this category.

Is Gemini 3 Pro good for coding and programming?

Gemini 3 Pro ranks #16 out of 122 models in coding and programming benchmarks with an average score of 62.8. There are stronger options in this category.

Is Gemini 3 Pro good for mathematics?

Gemini 3 Pro has visible benchmark coverage in mathematics, but BenchLM does not currently assign it a global category rank there.

Is Gemini 3 Pro good for reasoning and logic?

Gemini 3 Pro has visible benchmark coverage in reasoning and logic, but BenchLM does not currently assign it a global category rank there.

Is Gemini 3 Pro good for agentic tool use and computer tasks?

Gemini 3 Pro ranks #13 out of 119 models in agentic tool use and computer tasks benchmarks with an average score of 59.7. There are stronger options in this category.

Is Gemini 3 Pro good for multimodal and grounded tasks?

Gemini 3 Pro ranks #9 out of 29 models in multimodal and grounded tasks benchmarks with an average score of 73.2. It is among the top performers in this category.

Is Gemini 3 Pro good for instruction following?

Gemini 3 Pro has visible benchmark coverage in instruction following, but BenchLM does not currently assign it a global category rank there.

Is Gemini 3 Pro good for multilingual tasks?

Gemini 3 Pro has visible benchmark coverage in multilingual tasks, but BenchLM does not currently assign it a global category rank there.

Which sibling models are related to Gemini 3 Pro?

Gemini 3 Pro belongs to the Gemini 3 Pro family. Related variants on BenchLM include Gemini 3 Pro Deep Think.

Does Gemini 3 Pro have full benchmark coverage on BenchLM?

Not yet. Gemini 3 Pro currently has 27 published benchmark scores out of the 321 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of Gemini 3 Pro?

Gemini 3 Pro has a published context window of 2M, which determines how much text it can process in a single interaction.

Last updated: July 20, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Choose with this week’s evidence

Join 2,000+ readers for ranking moves, new releases, pricing changes, and the evidence behind them.

Free. One email per week.