Gemini 3.6 Flash places in the top three on MLE-bench.
Google DeepMind Jul 2026 Closed weights Mid-size
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LMArena (Elo) | 1,483 | #20 of 333 | Official leaderboard | |
| SEA-HELM | 76.4 | #5 of 62 | Official leaderboard | |
| MRCR v2 | 91.8% | #4 of 27 | Model card | |
| FACTS Multimodal | 47.2% | #6 of 39 | Official leaderboard | |
| SimpleQA Verified | 66.2% | #13 of 84 | Epoch AI | |
| FACTS Benchmark Suite | 63.1% | #8 of 40 | Official leaderboard | |
| FACTS Parametric | 73.3% | #10 of 45 | Official leaderboard | |
| Global MMLU Lite | 94.3% | #10 of 43 | Official leaderboard | |
| Extended NYT Connections | 89% | #25 of 97 | Official leaderboard | |
| FACTS Parametric V2 | 73.9% | #6 of 20 | Official leaderboard | |
| Creative Writing v3 | 72.5 | #40 of 120 | Official leaderboard | |
| BullshitBench V2 | 39% | #58 of 122 | Official leaderboard | |
| Document Arena | 1,453 | #24 of 37 | Official leaderboard | |
| BullshitBench | 18.2% | #75 of 107 | Official leaderboard | |
| FACTS Search V2 | 68.2% | #18 of 21 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GPQA Diamond | 94.1% | #11 of 219 | Epoch AI | |
| DTBench | 92.5% | #21 of 178 | Official leaderboard | |
| OTIS Mock AIME 2024-2025 | 94.2% | #47 of 202 | Epoch AI | |
| MathArena (Apex, IMO, USAMO) | 26% | #12 of 47 | Official leaderboard | |
| ARC-AGI | 91.2% | #28 of 92 | Official leaderboard | |
| CritPt | 10.6% | #56 of 160 | Epoch AI | |
| ProofBench | 36% | #21 of 59 | Official leaderboard | |
| ARC-AGI-2 | 60.4% | #36 of 95 | Official leaderboard | |
| AIME 2026 | 96.7% | #12 of 31 | Official leaderboard | |
| FrontierMath Tiers 1-3 v2 | 59% | #43 of 89 | Epoch AI | |
| HMMT 2026 | 89.4% | #16 of 31 | Official leaderboard | |
| FrontierMath Tier 4 v2 | 21.9% | #48 of 71 | Epoch AI | |
| ForecastBench | 65.3 | #76 of 86 | Official leaderboard | |
| xbench-ScienceQA | 65.6% | #27 of 29 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MLE-bench | 63.9% | #3 of 18 | Model card | |
| AndroidBench | 75.6% | #10 of 32 | Official leaderboard | |
| WebDev Arena | 1,537 | #33 of 105 | Epoch AI | |
| WeirdML | 56.1% | #38 of 112 | Official leaderboard | |
| Code Arena | 1,538 | #45 of 124 | Official leaderboard | |
| SWE-Bench Pro | 58.7% | #29 of 77 | Model card | |
| ProgramBench | 0.50% | #10 of 25 | Official leaderboard | |
| Vibe Code Bench | 64% | #52 of 115 | Official leaderboard | |
| Image-to-WebDev Arena | 1,545 | #27 of 55 | Official leaderboard | |
| Code Migration | 30.9% | #41 of 78 | Official leaderboard | |
| FrontierCode | 34.4% | #30 of 46 | Epoch AI | |
| ALE-Bench | 716 | #81 of 121 | Epoch AI | |
| DeepSWE | 46.7% | #23 of 28 | Official leaderboard | |
| IOI v2 (Vals) | 35.1% | #39 of 45 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LLM Chess | 1,094 | #19 of 124 | Official leaderboard | |
| OSWorld-Verified | 83% | #6 of 37 | Model card | |
| Game Arena Coin Game | 438 | #7 of 35 | Official leaderboard | |
| Game Arena Chess | 1,238 | #8 of 39 | Official leaderboard | |
| Game Arena Dark Hex | 329 | #9 of 38 | Official leaderboard | |
| Game Arena Four in a Row | 496 | #10 of 38 | Official leaderboard | |
| Game Arena Reversi | 707 | #12 of 36 | Official leaderboard | |
| Game Arena Word Association | 288 | #12 of 36 | Official leaderboard | |
| Mystery Game Puzzles | 30% | #30 of 80 | Epoch AI | |
| Kaggle Game Arena | 325 | #15 of 39 | Official leaderboard | |
| Game Arena Poker | 2.79 | #12 of 31 | Official leaderboard | |
| Chess Puzzles | 42% | #33 of 83 | Epoch AI | |
| Game Arena Word Art | 620 | #14 of 35 | Official leaderboard | |
| Long-Horizon-Terminal-Bench (LHTB) | 39% | #10 of 25 | Official leaderboard | |
| Game Arena Checkers | 951 | #16 of 38 | Official leaderboard | |
| Game Arena Clobber | 295 | #18 of 36 | Official leaderboard | |
| Game Arena Ultimate Tic-Tac-Toe | 271 | #18 of 36 | Official leaderboard | |
| Game Arena Lines of Action | 480 | #19 of 36 | Official leaderboard | |
| Game Arena Nine Men's Morris | 594 | #19 of 36 | Official leaderboard | |
| Terminal-Bench 2.1 | 78% | #28 of 53 | Model card | |
| Game Arena Bargaining | 178 | #22 of 36 | Official leaderboard | |
| Game Arena Go | 719 | #22 of 36 | Official leaderboard | |
| Game Arena Dots and Boxes | 787 | #25 of 37 | Official leaderboard | |
| GAIA | 1.66% | #54 of 57 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LegalBench | 86.7% | #12 of 138 | Official leaderboard | |
| MedCode | 53.1% | #11 of 96 | Official leaderboard | |
| ProfBench | 57.4% | #9 of 72 | Official leaderboard | |
| TaxEval v2 | 74.9% | #28 of 129 | Official leaderboard | |
| Vals Finance Agent v2 | 56.3% | #20 of 80 | Official leaderboard | |
| MortgageTax | 67.6% | #23 of 86 | Official leaderboard | |
| Excel Modeling Benchmark | 65.4% | #25 of 76 | Official leaderboard | |
| CorpFin v2 | 63.3% | #46 of 119 | Official leaderboard | |
| Harvey LAB (Legal Agent Benchmark) | 3.33% | #43 of 80 | Official leaderboard | |
| PolicyBench | 82.6% | #28 of 50 | Official leaderboard | |
| MedScribe | 79.7% | #57 of 98 | Official leaderboard | |
| GDPval-AA v2 | 1,421 | #16 of 26 | Model card | |
| FrontierFinance | 46.3% | #10 of 16 | Official leaderboard | |
| APEX-Agents | 46.9% | #36 of 56 | Epoch AI | |
| Tax Agent Bench | 18% | #54 of 79 | Official leaderboard | |
| Legal Research Bench | 25% | #62 of 88 | Official leaderboard | |
| TaxCalcBench v2 | 20% | #19 of 26 | Official leaderboard | |
| EEBench | 37.4% | #20 of 26 | Official leaderboard | |
| CommerceAgentBench | 46.7% | #10 of 13 | Official leaderboard | |
| Public Benefits Bench | 56.6% | #41 of 50 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| CharXiv | 89.4% | #8 of 67 | Model card | |
| Vision Arena | 1,282 | #21 of 136 | Official leaderboard | |
| 1H-VideoQA | 82.2% | #4 of 20 | Official leaderboard | |
| ZeroBench | 18% | #17 of 66 | Official leaderboard | |
| Math Kangaroo 2025 (MathArena) | 87.8% | #6 of 22 | Official leaderboard | |
| ParseBench | 69.4% | #14 of 42 | Official leaderboard | |
| Humanity's Sixth Sense | 31.9% | #13 of 29 | Official leaderboard | |
| VoxelBench | 1,771 | #26 of 56 | Official leaderboard | |
| Blueprint-Bench 2 | 31.2% | #17 of 35 | Official leaderboard | |
| GDP.pdf | 14% | #33 of 39 | Epoch AI | |
| Furniture Assembly Benchmark | 23.3% | #30 of 35 | Epoch AI |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Human Pathogen Capabilities Test v2 (HPCT v2) | 65.8% | #10 of 71 | Official leaderboard | |
| DeepsecBench | 11.5 | #22 of 28 | Official leaderboard | |
| CyberBench (Vals) | 48.5% | #21 of 26 | Official leaderboard | |
| CyberBench v1.1 (Vals) | 45.4% | #37 of 43 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Conceptual Reasoning Index | 59.5 | #34 of 149 | Official leaderboard | |
| LMCA | 44.9% | #39 of 139 | Epoch AI | |
| SAGE (Vals) | 48.7% | #25 of 81 | Official leaderboard | |
| ACCoRD | 70.0 | #52 of 150 | Official leaderboard | |
| Vals Multimodal Index | 65.1% | #13 of 33 | Official leaderboard |
Google DeepMind's models in color, against every other model on the Epoch Capabilities Index, by release date.
The benchmarks the lab chose to headline, with the scores it gave.
Every result in the longer documents the lab published with it. 5 of them aren't in the launch post.