Other · model
Gemma 4 31B benchmark scores
As of 2026-11-21, Gemma 4 31B (Other) has measured scores on 28 benchmarks and estimated scores on 128 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 85.7% | measured |
| AA-HLE | 23.6% | measured |
| AA-SciCode | 45.5% | measured |
| Artificial Analysis Intelligence Index | 4.4% | measured |
| CritPt | 1.4% | measured |
| GDPval-AA | 5.9% | measured |
| GPQA Diamond | 84.4% | estimated ± 1.2 pp, high confidence |
| HLE | 26.5% | measured |
| IFBench | 73.8% | estimated ± 6.6 pp, medium confidence |
| AA-LCR | 69.7% | measured |
| LiveCodeBench v6 | 80.0% | measured |
| AA-Omniscience Accuracy | 19.9% | measured |
| SciCode | 45.5% | estimated ± 3.1 pp, high confidence |
| BrowseComp | 67.4% | estimated ± 5.7 pp, medium confidence |
| HLE w/ tools | 36.6% | estimated ± 5.3 pp, low confidence |
| IFEval | 93.0% | estimated ± 2.8 pp, high confidence |
| LongBench v2 | 58.4% | estimated ± 5.5 pp, medium confidence |
| τ²-bench results | 59.9% | measured |
| AA Coding Index | 43.4% | measured |
| AA-MMMU-Pro | 73.4% | measured |
| APEX-Agents | 3.8% | estimated ± 9.4 pp, low confidence |
| APEX-Agents-AA | 9.7% | estimated ± 9.5 pp, medium confidence |
| BioMysteryBench (human-difficult) | 20.9% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 63.9% | estimated ± 4.5 pp, medium confidence |
| SWE-bench Verified | 76.1% | estimated ± 5.0 pp, medium confidence |
| BFCL v4 | 64.1% | estimated ± 2.7 pp, high confidence |
| CyberGym | 50.7% | estimated ± 5.3 pp, low confidence |
| JobBench | 0.0% | estimated ± 8.7 pp, low confidence |
| SWE-bench Pro | 38.9% | estimated ± 6.0 pp, medium confidence |
| WideResearch | 44.2% | estimated ± 5.9 pp, low confidence |
| AA Agentic Index | 6.7% | measured |
| MMLU-Pro | 85.2% | measured |
| AA-IFBench | 75.6% | measured |
| BBH | 98.9% | estimated ± 2.8 pp, medium confidence |
| FailureSensorIQ | 66.6% | estimated ± 1.7 pp, high confidence |
| GPQA | 84.3% | measured |
| MMLU | 86.2% | estimated ± 0.9 pp, high confidence |
| Vals GPQA Diamond | 77.1% | estimated ± 2.9 pp, medium confidence |
| Vals LiveCodeBench | 0.1% | estimated ± 3.4 pp, low confidence |
| Vals MMLU-Pro | 77.8% | estimated ± 1.8 pp, medium confidence |
| MMMLU | 86.8% | estimated ± 0.7 pp, medium confidence |
| MMMU | 82.9% | estimated ± 0.8 pp, high confidence |
| Chess Puzzles | 30.8% | estimated ± 10.8 pp, low confidence |
| AA MMLU-Pro | 84.1% | estimated ± 0.8 pp, medium confidence |
| Gert Labs | 35.3% | measured |
| AA Harvey LAB v1.0 | 79.4% | estimated ± 1.2 pp, low confidence |
| AA Tau3 Banking | 19.9% | estimated ± 3.7 pp, medium confidence |
| ARC-AGI-1 | 28.0% | estimated ± 3.9 pp, medium confidence |
| ARC-AGI-2 | 16.9% | estimated ± 10.9 pp, low confidence |
| CursorBench 3.1 | 32.9% | estimated ± 3.4 pp, low confidence |
| CursorBench 3.2 | 28.9% | estimated ± 4.4 pp, medium confidence |
| EBR-bench | 21.6% | estimated ± 11.3 pp, low confidence |
| FrontierCode 1.1 Main | 15.9% | estimated ± 4.6 pp, medium confidence |
| FrontierSWE v2 | 11.2% | estimated ± 14.2 pp, low confidence |
| Furniture Assembly | 33.8% | estimated ± 12.4 pp, low confidence |
| MaxxBench | 43.1% | estimated ± 5.7 pp, low confidence |
| MazeBench | 0.4% | estimated ± 5.5 pp, low confidence |
| Mystery Game Puzzles | 17.2% | estimated ± 13.3 pp, low confidence |
| NYT Connections | 70.6% | measured |
| OfficeQA Pro | 50.3% | estimated ± 6.8 pp, low confidence |
| OSWorld-Verified | 54.4% | estimated ± 5.5 pp, medium confidence |
| ParseBench-1 | 62.4% | measured |
| Vals SWE-bench | 65.7% | estimated ± 5.5 pp, medium confidence |
| CWE-Bench-AA | 9.2% | estimated ± 10.1 pp, low confidence |
| AA Cyber Index | 4.9% | estimated ± 9.2 pp, low confidence |
| DeepsecBench-AA | 0.0% | estimated ± 8.9 pp, low confidence |
| HealthBench Professional (length-adjusted, Artificial Analysis) | 57.5% | estimated ± 4.4 pp, low confidence |
| HealthBench Professional (raw, Artificial Analysis) | 12.7% | estimated ± 2.1 pp, low confidence |
| AA ITBench | 37.3% | measured |
| Bug Hunt Bench | 18.5% | estimated ± 10.5 pp, low confidence |
| Chartography (tools) | 79.2% | estimated ± 4.1 pp, low confidence |
| CursorBench 4.0 | 31.8% | estimated ± 4.9 pp, medium confidence |
| CWE-bench v1 | 9.9% | estimated ± 6.4 pp, low confidence |
| DeepSWE | 47.2% | estimated ± 7.5 pp, low confidence |
| DRACO | 38.4% | estimated ± 2.0 pp, low confidence |
| FrontierCode 2.0 | 26.1% | estimated ± 9.3 pp, low confidence |
| FrontierPhysics | 1.2% | estimated ± 4.7 pp, medium confidence |
| HLE w/o tools | 19.5% | measured |
| PostTrainBench v1.1 | 27.0% | estimated ± 7.4 pp, low confidence |
| SpireClimb | 5.8% | estimated ± 7.2 pp, medium confidence |
| StealthMark | 29.5% | measured |
| SWE-bench Multimodal | 38.9% | estimated ± 5.6 pp, low confidence |
| Toolathlon-Verified | 22.4% | estimated ± 2.3 pp, low confidence |
| VulcanBench Frontier v4 | 75.7% | estimated ± 6.8 pp, low confidence |
| GDP.pdf | 17.0% | estimated ± 4.6 pp, medium confidence |
| ARC-AGI-3 | 0.0% | estimated ± 7.2 pp, low confidence |
| ARC-AGI-3 Provider Adapter | 2.7% | estimated ± 7.3 pp, low confidence |
| cursorBench41 | 44.8% | estimated ± 3.7 pp, medium confidence |
| SkillsBench official leaderboard (BenchFlow) | 52.8% | estimated ± 6.1 pp, medium confidence |
| BinaryAudit | 35.8% | estimated ± 12.1 pp, low confidence |
| OTelBench | 22.5% | estimated ± 6.6 pp, medium confidence |
| Vibe Code Bench | 27.7% | estimated ± 6.5 pp, low confidence |
| FACTS Multimodal | 42.3% | measured |
| Hermes Bench | 24.0% | estimated ± 2.9 pp, medium confidence |
| Hermes Index | 5.9% | estimated ± 8.7 pp, low confidence |
| ParseBench-2 | 67.0% | estimated ± 3.5 pp, high confidence |
| CharXiv | 84.6% | estimated ± 5.6 pp, low confidence |
| CharXiv w/o tools | 81.0% | estimated ± 1.2 pp, high confidence |
| ExploitBench | 62.4% | estimated ± 7.3 pp, low confidence |
| ExploitGym | 0.0% | estimated ± 10.6 pp, low confidence |
| AI2D_TEST | 89.4% | estimated ± 2.0 pp, medium confidence |
| C-Eval | 92.2% | estimated ± 0.9 pp, medium confidence |
| Claw-Eval | 52.7% | estimated ± 6.8 pp, low confidence |
| CountBench | 93.9% | estimated ± 2.9 pp, medium confidence |
| DynaMath | 84.6% | estimated ± 3.0 pp, medium confidence |
| ERQA | 56.4% | estimated ± 3.1 pp, high confidence |
| MathVision | 86.1% | estimated ± 3.0 pp, high confidence |
| MCP Atlas | 69.1% | estimated ± 6.8 pp, medium confidence |
| MLVU (M-Avg) | 85.4% | estimated ± 2.2 pp, high confidence |
| MMLU-Redux | 93.3% | estimated ± 0.6 pp, high confidence |
| MMMU-Pro | 76.9% | measured |
| NL2Repo | 29.1% | estimated ± 3.4 pp, medium confidence |
| QwenClawBench | 55.9% | estimated ± 6.0 pp, low confidence |
| RealWorldQA | 83.4% | estimated ± 2.8 pp, high confidence |
| ScreenSpot Pro | 74.7% | estimated ± 9.6 pp, low confidence |
| SimpleVQA | 65.0% | estimated ± 7.5 pp, medium confidence |
| SuperGPQA | 67.2% | estimated ± 7.2 pp, medium confidence |
| Toolathlon | 48.8% | estimated ± 5.1 pp, low confidence |
| V* | 100.0% | estimated ± 7.5 pp, medium confidence |
| Video-MME (w/o subtitle) | 83.9% | estimated ± 1.3 pp, high confidence |
| Video-MME (with subtitle) | 86.3% | estimated ± 3.0 pp, high confidence |
| VideoMMMU | 84.7% | estimated ± 2.4 pp, high confidence |
| We-Math | 81.7% | estimated ± 7.4 pp, low confidence |
| AA LiveCodeBench | 88.1% | estimated ± 4.7 pp, low confidence |
| HealthBench Hard | 12.3% | estimated ± 8.6 pp, low confidence |
| MedXpertQA (MM) | 67.7% | estimated ± 2.4 pp, high confidence |
| MedXpertQA (Text) | 32.8% | estimated ± 8.9 pp, low confidence |
| MMLU-Pro (Arcee) | 82.4% | estimated ± 4.0 pp, medium confidence |
| React Native Evals | 75.2% | measured |
| ResearchClawBench | 9.5% | estimated ± 1.9 pp, medium confidence |
| SWE-Rebench | 41.6% | measured |
| DAYJOB | 0.0% | estimated ± 3.2 pp, medium confidence |
| DAYJOB Healthcare | 0.0% | estimated ± 4.7 pp, medium confidence |
| BioMysteryBench (human-solvable) | 81.7% | estimated ± 0.6 pp, low confidence |
| HealthBench (raw) | 31.9% | estimated ± 6.0 pp, low confidence |
| HealthBench (length-adjusted) | 47.3% | estimated ± 3.4 pp, low confidence |
| HealthBench Professional | 28.3% | estimated ± 4.2 pp, low confidence |
| HealthBench Professional (raw) | 43.6% | estimated ± 4.5 pp, medium confidence |
| HLE-Verified | 15.3% | estimated ± 7.1 pp, low confidence |
| ScarfBench | 24.4% | estimated ± 9.1 pp, low confidence |
| BlenderWorld (text-only) | 64.8% | estimated ± 4.3 pp, low confidence |
| Cookie-Bench | 78.0% | estimated ± 4.8 pp, low confidence |
| STEMBench | 2.3% | estimated ± 0.2 pp, low confidence |
| SimpleQA | 18.0% | estimated ± 4.3 pp, medium confidence |
| BabyVision w/ Python | 88.8% | estimated ± 3.8 pp, low confidence |
| OpenHarmony Bench v1.0 | 51.0% | estimated ± 2.4 pp, high confidence |
| SEC-Bench Pro | 56.5% | estimated ± 5.6 pp, medium confidence |
| BabyVision | 58.2% | estimated ± 9.1 pp, low confidence |
| MMVU | 76.5% | estimated ± 2.7 pp, medium confidence |
| PerceptionBench | 59.1% | estimated ± 4.9 pp, low confidence |
| MRCRv2 | 66.4% | measured |
| MMLongBench-Doc | 60.9% | estimated ± 2.6 pp, low confidence |
| OCRBench V2 | 70.8% | estimated ± 3.3 pp, high confidence |
| BioBasicsBench | 33.9% | estimated ± 2.8 pp, low confidence |
| MMMU-Pro w/ Python | 76.8% | estimated ± 2.0 pp, high confidence |
| ExploitBench Internal Port | 0.0% | estimated ± 6.8 pp, low confidence |