Other · model
Gemma 4 12B benchmark scores
As of 2026-11-22, Gemma 4 12B (Other) has measured scores on 21 benchmarks and estimated scores on 100 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 75.3% | measured |
| AA-HLE | 15.7% | measured |
| AA-SciCode | 36.2% | estimated ± 3.4 pp, high confidence |
| AIME26 | 77.5% | measured |
| Artificial Analysis Intelligence Index | 10.7% | estimated ± 5.5 pp, low confidence |
| CritPt | 0.0% | measured |
| GDPval-AA | 0.0% | measured |
| GPQA Diamond | 78.8% | measured |
| HLE | 8.0% | estimated ± 6.5 pp, low confidence |
| IFBench | 71.5% | estimated ± 6.6 pp, medium confidence |
| AA-LCR | 63.7% | measured |
| LiveCodeBench v6 | 72.0% | measured |
| AA-Omniscience Accuracy | 15.7% | measured |
| SciCode | 36.8% | estimated ± 4.6 pp, high confidence |
| BrowseComp | 0.0% | estimated ± 9.2 pp, low confidence |
| HLE w/ tools | 33.8% | estimated ± 5.3 pp, low confidence |
| IFEval | 93.0% | estimated ± 2.8 pp, high confidence |
| LongBench v2 | 55.6% | estimated ± 5.5 pp, medium confidence |
| τ²-bench results | 36.3% | measured |
| AA Coding Index | 31.0% | measured |
| AA-MMMU-Pro | 69.7% | measured |
| APEX-Agents | 0.0% | estimated ± 9.4 pp, low confidence |
| APEX-Agents-AA | 6.0% | estimated ± 9.5 pp, medium confidence |
| BioMysteryBench (human-difficult) | 0.0% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 34.7% | estimated ± 9.6 pp, medium confidence |
| SWE-bench Verified | 57.5% | estimated ± 6.5 pp, medium confidence |
| BFCL v4 | 53.9% | estimated ± 2.7 pp, high confidence |
| CyberGym | 0.0% | estimated ± 6.0 pp, low confidence |
| JobBench | 0.0% | estimated ± 8.7 pp, low confidence |
| SWE-bench Pro | 22.3% | estimated ± 6.0 pp, medium confidence |
| WideResearch | 0.0% | estimated ± 7.8 pp, low confidence |
| AA Agentic Index | 0.0% | estimated ± 3.1 pp, high confidence |
| MMLU-Pro | 77.2% | measured |
| AA-IFBench | 73.5% | measured |
| BBH | 94.2% | estimated ± 2.8 pp, medium confidence |
| GPQA | 78.8% | measured |
| MMLU | 80.4% | estimated ± 0.9 pp, high confidence |
| Vals GPQA Diamond | 78.5% | estimated ± 2.1 pp, medium confidence |
| Vals LiveCodeBench | 66.1% | estimated ± 7.1 pp, medium confidence |
| Vals MMLU-Pro | 80.2% | estimated ± 1.6 pp, medium confidence |
| MATH-500 | 95.2% | estimated ± 3.1 pp, medium confidence |
| MMMLU | 83.4% | measured |
| MMMU | 79.3% | estimated ± 0.8 pp, high confidence |
| AIME 2025 | 73.9% | estimated ± 1.8 pp, medium confidence |
| AA MMLU-Pro | 83.4% | estimated ± 0.8 pp, medium confidence |
| Gert Labs | 26.9% | estimated ± 10.0 pp, low confidence |
| AA Tau3 Banking | 0.0% | estimated ± 4.1 pp, medium confidence |
| ARC-AGI-1 | 13.6% | estimated ± 3.9 pp, medium confidence |
| ARC-AGI-2 | 0.0% | estimated ± 10.9 pp, low confidence |
| FrontierSWE v2 | 0.2% | estimated ± 14.3 pp, low confidence |
| MazeBench | 0.4% | estimated ± 5.5 pp, low confidence |
| OfficeQA Pro | 50.3% | estimated ± 6.8 pp, low confidence |
| OSWorld-Verified | 55.4% | estimated ± 8.4 pp, low confidence |
| ParseBench-1 | 53.5% | estimated ± 6.2 pp, medium confidence |
| Vals SWE-bench | 54.8% | estimated ± 5.5 pp, medium confidence |
| CWE-Bench-AA | 0.0% | estimated ± 10.1 pp, low confidence |
| DeepsecBench-AA | 0.0% | estimated ± 8.9 pp, low confidence |
| HealthBench Professional (raw, Artificial Analysis) | 1.3% | estimated ± 2.1 pp, low confidence |
| AA ITBench | 35.7% | estimated ± 5.1 pp, low confidence |
| Chartography (tools) | 76.8% | estimated ± 4.1 pp, low confidence |
| CWE-bench v1 | 0.0% | estimated ± 6.4 pp, low confidence |
| DRACO | 0.0% | estimated ± 2.0 pp, low confidence |
| HLE w/o tools | 5.2% | measured |
| Toolathlon-Verified | 0.0% | estimated ± 3.1 pp, medium 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 |
| Vibe Code Bench | 9.8% | estimated ± 10.0 pp, low confidence |
| Hermes Bench | 0.0% | estimated ± 2.9 pp, medium confidence |
| Hermes Index | 0.0% | estimated ± 8.7 pp, low confidence |
| ParseBench-2 | 57.1% | estimated ± 5.0 pp, medium confidence |
| CharXiv | 76.4% | estimated ± 3.8 pp, high confidence |
| CharXiv w/o tools | 75.3% | estimated ± 1.2 pp, high confidence |
| AI2D_TEST | 90.1% | estimated ± 1.0 pp, medium confidence |
| C-Eval | 89.3% | estimated ± 1.0 pp, medium confidence |
| CountBench | 91.7% | estimated ± 1.3 pp, medium confidence |
| DynaMath | 66.9% | estimated ± 3.0 pp, low confidence |
| ERQA | 43.3% | estimated ± 3.1 pp, medium confidence |
| HMMT Feb 2026 | 55.9% | estimated ± 4.0 pp, high confidence |
| MathVision | 79.7% | measured |
| MCP Atlas | 56.5% | estimated ± 10.7 pp, low confidence |
| MLVU (M-Avg) | 83.1% | estimated ± 1.5 pp, high confidence |
| MMLU-Redux | 92.2% | estimated ± 0.5 pp, medium confidence |
| MMMU-Pro | 69.1% | measured |
| NL2Repo | 20.0% | estimated ± 3.4 pp, medium confidence |
| OmniDocBench 1.5 | 87.6% | estimated ± 2.2 pp, high confidence |
| QwenClawBench | 55.9% | estimated ± 6.0 pp, low confidence |
| RealWorldQA | 79.6% | estimated ± 2.3 pp, high confidence |
| ScreenSpot Pro | 53.6% | estimated ± 9.3 pp, medium confidence |
| SimpleVQA | 64.0% | estimated ± 7.3 pp, medium confidence |
| SuperGPQA | 59.7% | estimated ± 8.5 pp, low confidence |
| Toolathlon | 27.0% | estimated ± 6.3 pp, low confidence |
| V* | 45.6% | estimated ± 7.5 pp, low confidence |
| Video-MME (w/o subtitle) | 79.5% | estimated ± 1.3 pp, high confidence |
| Video-MME (with subtitle) | 83.2% | estimated ± 1.9 pp, high confidence |
| VideoMMMU | 78.6% | estimated ± 2.4 pp, high confidence |
| We-Math | 69.6% | estimated ± 2.1 pp, medium confidence |
| AA LiveCodeBench | 84.0% | estimated ± 4.7 pp, low confidence |
| HealthBench Hard | 0.0% | estimated ± 8.6 pp, low confidence |
| MedXpertQA (MM) | 48.7% | measured |
| MMLU-Pro (Arcee) | 82.8% | estimated ± 3.4 pp, medium confidence |
| React Native Evals | 71.8% | estimated ± 3.2 pp, high confidence |
| ResearchClawBench | 11.1% | estimated ± 2.2 pp, low confidence |
| BioMysteryBench (human-solvable) | 81.7% | estimated ± 0.6 pp, low confidence |
| HealthBench (raw) | 22.4% | estimated ± 6.0 pp, low confidence |
| HealthBench (length-adjusted) | 44.3% | estimated ± 3.4 pp, low confidence |
| HealthBench Professional | 58.1% | estimated ± 4.5 pp, low confidence |
| HealthBench Professional (raw) | 34.9% | estimated ± 4.5 pp, medium confidence |
| ScarfBench | 24.4% | estimated ± 9.1 pp, low confidence |
| STEMBench | 2.3% | estimated ± 0.2 pp, low confidence |
| SimpleQA | 12.2% | estimated ± 4.3 pp, medium confidence |
| BabyVision w/ Python | 88.8% | estimated ± 3.8 pp, low confidence |
| IMOAnswerBench | 64.9% | estimated ± 10.7 pp, low confidence |
| BabyVision | 32.7% | estimated ± 3.7 pp, high confidence |
| MMVU | 71.5% | estimated ± 2.7 pp, medium confidence |
| PerceptionBench | 54.7% | estimated ± 4.9 pp, low confidence |
| MRCRv2 | 43.4% | measured |
| MMLongBench-Doc | 60.6% | estimated ± 2.6 pp, low confidence |
| OCRBench V2 | 68.6% | estimated ± 3.3 pp, high confidence |
| BioBasicsBench | 33.9% | estimated ± 10.7 pp, low confidence |
| MMMU-Pro w/ Python | 73.5% | estimated ± 2.0 pp, high confidence |