Gemma 4 31B IT ranks first on 3 of the 77 benchmarks it has scores on.
Google DeepMind Apr 2026 Open weights Open model
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 | |
|---|---|---|---|---|
| IFEval | 98.9% | #1 of 126 | Model card | |
| RULER | 96.4% | #2 of 38 | Model card | |
| BIG-Bench Extra Hard | 74.4% | #1 of 11 | Model card | |
| HELM Arabic Enterprise | 73.8 | #3 of 23 | Official leaderboard | |
| HELM Arabic | 82.3 | #4 of 28 | Official leaderboard | |
| SEA-HELM | 76.2 | #9 of 62 | Official leaderboard | |
| MMLU-Pro | 85.2% | #22 of 151 | Model card | |
| LMArena (Elo) | 1,452 | #60 of 333 | Official leaderboard | |
| AgentIF (2026-10 setting) | 64.8% | #6 of 24 | Official leaderboard | |
| IFBench | 76% | #21 of 70 | Model card | |
| MMMLU | 88.4% | #19 of 62 | Model card | |
| EQ-Bench 3 | 1,464 | #25 of 78 | Official leaderboard | |
| FACTS Benchmark Suite | 56.2% | #13 of 40 | Official leaderboard | |
| Judgemark v4 | 72.3 | #13 of 40 | Official leaderboard | |
| FACTS Grounding v2 | 75.8% | #5 of 15 | Official leaderboard | |
| Vectara Hallucination Leaderboard (2025 dataset) | 7.40% | #37 of 104 | Official leaderboard | |
| Wer wird Millionär? LLM Benchmark | 26311.0 | #28 of 76 | Official leaderboard | |
| GraphWalks | 82.3% | #8 of 21 | Model card | |
| Global MMLU Lite | 90.7% | #20 of 43 | Official leaderboard | |
| FACTS Search | 68.9% | #15 of 31 | Official leaderboard | |
| Creative Writing v3 | 68.0 | #61 of 120 | Official leaderboard | |
| Extended NYT Connections | 70.6% | #51 of 97 | Official leaderboard | |
| Humanity's Last Exam | 26.5% | #82 of 147 | Model card | |
| Longform Creative Writing (EQ-Bench) | 56.5 | #74 of 119 | Official leaderboard | |
| MRCR v2 | 66.4% | #17 of 27 | Model card | |
| BullshitBench | 21.8% | #68 of 107 | Official leaderboard | |
| BullshitBench V2 | 25% | #78 of 122 | Official leaderboard | |
| Document Arena | 1,445 | #29 of 37 | Official leaderboard | |
| EQ-Bench 4 | 1,110 | #26 of 32 | Official leaderboard | |
| LOFT | 79.5% | #5 of 6 | Model card | |
| SimpleQA Verified | 10.7% | #47 of 56 | Official leaderboard | |
| FACTS Multimodal | 24.7% | #33 of 39 | Official leaderboard | |
| FACTS Parametric | 7.79% | #39 of 45 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ForecastBench | 61.3 | #28 of 86 | Official leaderboard | |
| DTBench | 71.1% | #77 of 178 | Official leaderboard | |
| GPQA Diamond | 75.8% | #101 of 219 | Epoch AI | |
| OTIS Mock AIME 2024-2025 | 73.3% | #95 of 202 | Epoch AI | |
| CritPt | 1.43% | #88 of 160 | Epoch AI | |
| Surface Evolver Bench | 30.6% | #21 of 26 | Epoch AI | |
| AIME 2026 | 89.2% | #28 of 34 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| StructEval | 77.5% | #5 of 28 | Official leaderboard | |
| Codeforces rating | 2,150 | #14 of 41 | Model card | |
| WeirdML | 52.3% | #46 of 130 | Epoch AI | |
| LiveCodeBench v6 | 80% | #29 of 76 | Model card | |
| ALE-Bench | 926 | #60 of 121 | Epoch AI | |
| SciCode | 43% | #24 of 47 | Model card | |
| Code Arena | 1,369 | #91 of 124 | Official leaderboard | |
| WebDev Arena | 1,364 | #78 of 105 | Epoch AI | |
| AndroidBench | 37.1% | #31 of 32 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Terminal-Bench Hard | 36% | #1 of 11 | Model card | |
| PACT | 1,557 | #8 of 26 | Official leaderboard | |
| tau2-bench | 76.9% | #41 of 69 | Model card | |
| PinchBench | 66.5% | #30 of 48 | Official leaderboard | |
| SnakeBench | 18 | #149 of 183 | Official leaderboard | |
| PinchBench 2.0 | 52.7% | #48 of 57 | Official leaderboard | |
| Pencil Puzzle Bench | 0% | #51 of 53 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vals Finance Agent | 50.8% | #27 of 49 | Official leaderboard | |
| MedXpertQA-MM | 61.3% | #7 of 12 | Model card | |
| ProfBench | 44.4% | #47 of 72 | Official leaderboard | |
| MortgageTax | 61.4% | #57 of 86 | Official leaderboard | |
| RadLE 2.0 | 336.0 | #9 of 12 | Official leaderboard | |
| CaseLaw v2 | 52.6% | #44 of 53 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| InfographicVQA | 92% | #2 of 22 | Model card | |
| Vision Arena | 1,262 | #38 of 136 | Official leaderboard | |
| VideoMME v2 (with subtitles) | 28.2% | #9 of 31 | Official leaderboard | |
| MATH-Vision (MathVision) | 85.6% | #14 of 42 | Model card | |
| MMMU-Pro | 76.9% | #26 of 72 | Model card | |
| ParseBench | 62.4% | #23 of 42 | Official leaderboard | |
| OmniDocBench v1.5 | 0.13 | #9 of 13 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Human Pathogen Capabilities Test v2 (HPCT v2) | 43.8% | #51 of 71 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SAGE (Vals) | 55% | #2 of 81 | Official leaderboard | |
| Decision Index | 67.6 | #1 of 3 | Official leaderboard | |
| LMCA | 39.3% | #53 of 139 | Epoch AI | |
| Political Science LLM Benchmark | 66.4% | #13 of 29 | Official leaderboard | |
| Conceptual Reasoning Index | 50.2 | #67 of 149 | Official leaderboard | |
| ACCoRD | 62.0 | #81 of 150 | Official leaderboard | |
| SAGE v1.1 (Vals) | 82.8% | #23 of 28 | Official leaderboard | |
| LiveBench | 38.7% | #28 of 28 | 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. 10 of them aren't in the launch post.