GoogleGemini 3 Pro
Reasoning effort
Gemini 3 Pro is a Google proprietary reasoning model, released 18 Nov 2025. Its best configuration ranks #109 of 785 on the BenchLeader Index at 60.4 ±3.4, in the upper half. The ± is the point: 206 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in human preference (68) and lowest in maths (54). It has been measured at 4 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Oct 2026.
- Blended price
- –
- Output speed
- –
- First answer
- –
- Context
- –
- Full answer
- –
- Cost per run
- –
- Released
- 18 Nov 2025
- Overall index60
- Reasoning66
- Coding58
- Agents & tools62
- Maths54
- Knowledge56
- Instruction following60
- Human preference68
- Multimodal66
Versions
Google has shipped 10 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Gemini 3.1 Pro | 19 Feb 2026 | 60.6 | #106 |
| Gemini 3 Prothis page | 18 Nov 2025 | 60.4 | #109 |
| Gemini 2.5 Pro | 25 Mar 2025 | 53.9 | #264 |
| Gemini 2.0 Pro | 5 Feb 2025 | 47.0 | #464 |
| Gemini 1.5 Pro 002 | 24 Sept 2024 | 41.6 | #623 |
| Gemini 1.5 Pro | 15 May 2024 | 46.1 | #495 |
| Gemini 1.5 Pro 001 | 14 May 2024 | – | – |
| Gemini 1.0 Pro 001 | 13 Dec 2023 | – | – |
| Gemini 1.0 Pro | 6 Dec 2023 | – | – |
| Gemini 1.0 Pro 002 | – | – | – |
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| low | 53.8 | #271 | – | – | $0.0044 | Agents & tools 64 · Composite 54 · Instruction following 51 · Knowledge 62 · Long context 62 · Reasoning 41 |
| high | 60.0 | #118 | – | – | $0.0044 | Agents & tools 59 · Coding 61 · Composite 61 · Instruction following 69 · Knowledge 63 · Long context 63 · Maths 63 · Multimodal 63 · Reasoning 56 |
| thinking | – | – | – | – | – | |
| not statedbest | 60.4 | #109 | – | – | – | Agents & tools 62 · Coding 58 · Human preference 68 · Instruction following 60 · Knowledge 56 · Maths 54 · Multimodal 66 · Reasoning 66 |
How its index has moved
8 Sept 2026 to 23 Oct 2026- not stated
- high
- low
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Metadata
| Benchmark | high | Source | Trend |
|---|---|---|---|
| AA Openness Indexnot in index | 5.6#38 | Artificial Analysis |
Reasoning
| Benchmark | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamondeffort not stated | – | – | 92.6%#27 | Epoch AI Benchmarking Hub | |
| Humanity's Last Exameffort not stated | – | – | 37.5%#8 | Scale AI / CAIS | |
| SimpleBencheffort not stated | – | – | 76.4%#7 | SimpleBench | |
| Arena Hard Promptseffort not stated | – | – | 1502#29 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 88.7%#88 | 90.8%#55 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 29.5%#156 | 39.7%#89 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | 91.7%#18 | – | Vals AI | |
| Kagi LLM Benchmarkeffort not stated | – | – | 80.1%#9 | Kagi LLM Benchmark | |
| GPQA Diamond (HELM)not in indexeffort not stated | – | – | 80.3%#1 | HELM Capabilities | |
| ARC-AGI-1effort not stated | – | – | 75.0%#133 | ARC Prize | |
| ARC-AGI-2effort not stated | – | – | 54.0%#103 | ARC Prize | |
| CritPt | 0.0%#315 | 9.1%#128 | – | Artificial Analysis | |
| SciPredictnot in indexeffort not stated | – | – | 25.3%#1 | Scale AI SEAL | |
| CL-benchnot in indexeffort not stated | – | – | 15.8%#17 | Epoch AI Benchmarking Hub | |
| ForecastBenchnot in indexeffort not stated | – | – | 65.3%#14 | ForecastBench |
Coding
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| SWE-bench Verified (Epoch)effort not stated | – | 72.9%#21 | Epoch AI Benchmarking Hub | |
| WeirdMLeffort not stated | – | 69.9%#34 | WeirdML | |
| GSO-Bencheffort not stated | – | 18.6%#17 | GSO-Bench | |
| Arena Codingeffort not stated | – | 1517#43 | Arena (LMArena) | |
| Arena WebDeveffort not stated | – | 1440#75 | Arena (LMArena) | |
| LiveCodeBencheffort not stated | – | 86.4%#22 | Vals AI | |
| SWE-bench (Vals)not in index | 76.4%#35 | – | Vals AI | |
| SWE-Bench Proeffort not stated | – | 43.3%#8 | Scale AI SEAL | |
| SWE-bench Verified (bash only) | 69.6%#13 | 74.2%#6 | SWE-bench | |
| SWE-bench Verified (any scaffold)not in indexeffort not stated | – | 77.4%#2 | SWE-bench | |
| Vibe Code Bench v1.1not in index | 14.3%#90 | – | Vals AI | |
| SWE-Bench Pro (private)not in indexeffort not stated | – | 43.3%#8 | Scale AI SEAL | |
| ALE-Benchnot in indexeffort not stated | – | 1176.8#45 | Epoch AI Benchmarking Hub | |
| AlgoTunenot in indexeffort not stated | – | 1.8#4 | AlgoTune |
Agents & tools
| Benchmark | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| Terminal-Bencheffort not stated | – | – | 69.4%#7 | Terminal-Bench | |
| GDPvaleffort not stated | – | – | 40.3%#5 | OpenAI | |
| Remote Labor Indexeffort not stated | – | – | 1.3%#15 | Scale AI / CAIS | |
| Terminal-Bench Hard | 34.1%#89 | 41.7%#50 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 68.1%#161 | 87.1%#80 | – | Artificial Analysis | |
| MCP Atlaseffort not stated | – | – | 70.3%#20 | Scale AI SEAL | |
| BFCL Overalleffort not stated | – | – | 72.5%#3 | Berkeley Function Calling Leaderboard | |
| τ²-bench | – | 82.5%#5 | – | τ²-bench | |
| τ³-bench Bankingnot in index | – | 18.0%#19 | – | τ²-bench | |
| Poker Agentnot in indexeffort not stated | – | – | 1078.9%#6 | Vals AI | |
| Terminal-Bench 2.0 (Vals)not in index | – | 55.1%#15 | – | Vals AI | |
| BALROGnot in indexeffort not stated | – | – | 58.1%#4 | BALROG | |
| DeepResearch Benchnot in index | 46.3%#28 | – | – | DeepResearch Bench | |
| METR Time Horizonsnot in indexeffort not stated | – | – | 71.0%#8 | METR | |
| Vending-Bench 2not in indexeffort not stated | – | – | 5478.2#26 | Andon Labs |
Maths
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| OTIS Mock AIMEeffort not stated | – | 91.4%#68 | Epoch AI Benchmarking Hub | |
| ProofBencheffort not stated | – | 20.0%#51 | Vals AI | |
| AIME (Vals) | 96.7%#4 | – | Vals AI | |
| MGSM | 93.9%#7 | – | Vals AI | |
| AIME 2026effort not stated | – | 91.7%#24 | MathArena | |
| HMMT February 2026effort not stated | – | 86.4%#20 | MathArena | |
| MathArena Apexeffort not statedpossible contamination | – | 23.4%#14 | MathArena | |
| Omni-MATH (HELM)effort not stated | – | 55.6%#13 | HELM Capabilities | |
| MATH 500not in index | 96.4%#1 | – | Vals AI | |
| Arena Mathseffort not stated | – | 1477#42 | Arena (LMArena) |
Knowledge
| Benchmark | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| AA-Omniscience | 1.8#122 | 15.3#81 | – | Artificial Analysis | |
| MMLU-Proeffort not stated | – | – | 90.1%#7 | Vals AI | |
| LegalBench | – | 87.0%#7 | – | Vals AI | |
| CorpFin | – | 63.7%#42 | – | Vals AI | |
| TaxEval | – | 72.6%#60 | – | Vals AI | |
| MedQA | – | 96.0%#8 | – | Vals AI | |
| PRBench Financeeffort not stated | – | – | 39.2%#28 | Scale AI SEAL | |
| PRBench Legaleffort not stated | – | – | 40.6%#31 | Scale AI SEAL | |
| MultiNRCeffort not statedpossible contamination | – | – | 59.0%#6 | Scale AI SEAL | |
| MMLU-Pro (HELM)not in indexeffort not stated | – | – | 90.3%#1 | HELM Capabilities | |
| CaseLaw v2not in index | – | 53.1%#41 | – | Vals AI | |
| MedCodenot in index | – | 52.2%#14 | – | Vals AI | |
| MedScribenot in index | – | 72.0%#87 | – | Vals AI | |
| AA-Omniscience: accuracynot in index | 48.3%#71 | 55.8%#36 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 10.0%#421 | 8.6%#456 | – | Artificial Analysis | |
| Global-MMLU-Lite (AA)not in index | – | 92.2%#3 | – | Artificial Analysis |
Instruction following
| Benchmark | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| IFBench | 49.7%#177 | 70.4%#67 | – | Artificial Analysis | |
| MultiChallengeeffort not stated | – | – | 65.7%#9 | Scale AI SEAL | |
| TutorBencheffort not stated | – | – | 53.7%#9 | Scale AI SEAL | |
| IFEval (HELM)effort not stated | – | – | 87.6%#13 | HELM Capabilities | |
| WildBench (HELM)effort not stated | – | – | 85.9%#5 | HELM Capabilities | |
| Arena Instruction Followingnot in indexeffort not stated | – | – | 1473#33 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1486#20 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1485#10 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1496#14 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1491#26 | Arena (LMArena) |
Multimodal
| Benchmark | high | thinking | not stated | Source | Trend |
|---|---|---|---|---|---|
| Arena Visioneffort not stated | – | – | 1305#16 | Arena (LMArena) | |
| MMMU-Pro | 80.2%#51 | – | – | Artificial Analysis | |
| VISTAeffort not stated | – | – | 51.5%#7 | Scale AI SEAL | |
| MMMU-Pro (official)not in indexeffort not stated | – | – | 81.0%#1 | MMMU | |
| MMMU-Pro (Vals)not in index | 87.5%#14 | – | – | Vals AI | |
| MortgageTaxnot in index | 69.1%#7 | – | – | Vals AI | |
| SAGEnot in index | 47.6%#31 | – | – | Vals AI | |
| AudioMCnot in index | – | 54.6%#5 | – | Scale AI SEAL | |
| AudioMC (text output)not in index | – | 54.6%#5 | – | Scale AI SEAL | |
| VTBnot in indexeffort not stated | – | – | 26.9%#5 | Scale AI SEAL | |
| GeoBenchnot in indexeffort not stated | – | – | 84.0%#3 | GeoBench | |
| VPCTnot in indexeffort not stated | – | – | 91.0%#1 | Epoch AI Benchmarking Hub |
Long context
| Benchmark | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| AA-LCR | 74.0%#168 | 76.0%#148 | – | Artificial Analysis | |
| Arena Documentnot in indexeffort not stated | – | – | 1451#31 | Arena (LMArena) |
Composite
| Benchmark | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | 153.1#47 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2 | 22.3#202 | 28.0#139 | – | Artificial Analysis | |
| HELM Capabilities meannot in indexeffort not stated | – | – | 79.9%#5 | HELM Capabilities |
Safety & honesty
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| FORTRESSnot in index | 41.7%#18 | Scale AI SEAL | |
| MASKnot in indexpossible contamination | 42.6%#62 | Scale AI SEAL | |
| PropensityBenchnot in index | 52.9%#7 | Scale AI SEAL |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
More from Google
Compare Gemini 3 Pro with…
See also
Questions people ask
- What are Gemini 3 Pro's benchmark scores?
- As of 24 Oct 2026, Gemini 3 Pro scores 60.4 ±3.4 on the BenchLeader Index, strongest in human preference (68) and weakest in maths (54). That rests on 38 independent benchmark results, and every score on this page links to its publisher.
- How does Gemini 3 Pro rank?
- Gemini 3 Pro is #109 of 785 ranked configurations on 24 Oct 2026. 206 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- Is Gemini 3 Pro better than GPT-5.6 Sol?
- As of 24 Oct 2026, GPT-5.6 Sol (max) is the better pick for most workloads: it leads on quality (68.0 vs 60.4 on the BenchLeader Index). The head-to-head page has the benchmark-by-benchmark detail.
- When was Gemini 3 Pro released?
- Google released Gemini 3 Pro on 18 Nov 2025. BenchLeader has results for it at 4 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Gemini 3 Pro: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/gemini-3-pro.