Other · model
Gemma 4 12B benchmark scores
As of 2026-10-07, Gemma 4 12B (Other) has measured scores on 21 benchmarks and estimated scores on 92 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 75.3% | measured |
| AA-HLE | 15.7% | measured |
| AA-SciCode | 36.3% | estimated ± 3.2 pp, high confidence |
| AIME26 | 77.5% | measured |
| Artificial Analysis Intelligence Index | 14.2% | measured |
| CritPt | 0.0% | measured |
| GDPval-AA | 0.0% | measured |
| GPQA Diamond | 78.8% | measured |
| HLE | 8.2% | estimated ± 5.7 pp, low confidence |
| IFBench | 72.0% | estimated ± 6.9 pp, medium confidence |
| AA-LCR | 63.7% | measured |
| LiveCodeBench v6 | 72.0% | measured |
| AA-Omniscience Accuracy | 15.6% | measured |
| SciCode | 36.6% | estimated ± 4.7 pp, high confidence |
| BrowseComp | 0.0% | estimated ± 10.1 pp, low confidence |
| HLE w/ tools | 31.6% | estimated ± 6.1 pp, low confidence |
| IFEval | 93.6% | estimated ± 1.3 pp, high confidence |
| LongBench v2 | 61.5% | estimated ± 5.9 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 ± 10.0 pp, low confidence |
| APEX-Agents-AA | 4.7% | estimated ± 9.1 pp, medium confidence |
| BioMysteryBench (human-difficult) | 22.7% | estimated ± 5.9 pp, low confidence |
| DeepSearchQA | 34.5% | estimated ± 8.4 pp, medium confidence |
| SWE-bench Verified | 56.8% | estimated ± 5.5 pp, medium confidence |
| AutomationBench | 28.4% | estimated ± 9.7 pp, low confidence |
| BFCL v4 | 53.8% | estimated ± 2.6 pp, high confidence |
| CyberGym | 0.0% | estimated ± 6.2 pp, low confidence |
| JobBench | 0.0% | estimated ± 8.6 pp, low confidence |
| SWE-bench Pro | 20.1% | estimated ± 4.7 pp, high confidence |
| WideResearch | 41.6% | estimated ± 9.1 pp, low confidence |
| AA Agentic Index | 2.7% | estimated ± 2.8 pp, high confidence |
| MMLU-Pro | 77.2% | measured |
| GPQA | 78.8% | measured |
| AA-IFBench | 73.5% | measured |
| Gert Labs | 36.0% | estimated ± 8.9 pp, medium confidence |
| AA-AnalystAgent | 0.0% | estimated ± 12.0 pp, low confidence |
| AA EnterpriseOps-Gym | 1.4% | estimated ± 3.6 pp, low confidence |
| ARC-AGI-1 | 67.8% | estimated ± 2.7 pp, medium confidence |
| ARC-AGI-2 | 28.4% | estimated ± 10.1 pp, low confidence |
| CursorBench 3.1 | 47.9% | estimated ± 4.8 pp, low confidence |
| CursorBench 3.2 | 18.9% | estimated ± 4.4 pp, medium confidence |
| FrontierCode 1.1 Main | 10.5% | estimated ± 3.8 pp, medium confidence |
| FrontierSWE v2 | 0.2% | estimated ± 14.8 pp, low confidence |
| OfficeQA Pro | 50.3% | estimated ± 7.2 pp, low confidence |
| OSWorld-Verified | 55.6% | estimated ± 8.7 pp, low confidence |
| Vals GPQA Diamond | 77.7% | estimated ± 2.0 pp, medium confidence |
| Vals MMLU-Pro | 64.9% | estimated ± 1.2 pp, medium confidence |
| Vals SWE-bench | 56.2% | estimated ± 5.4 pp, low confidence |
| AA AutomationBench | 3.7% | estimated ± 7.0 pp, low confidence |
| GDP.pdf | 9.4% | estimated ± 5.4 pp, low confidence |
| AA ITBench | 99.0% | estimated ± 3.6 pp, low confidence |
| AA Tau3 Banking | 22.0% | estimated ± 5.8 pp, low confidence |
| CursorBench 4.0 | 7.7% | estimated ± 1.6 pp, medium confidence |
| CWE-bench v1 | 0.0% | estimated ± 6.9 pp, low confidence |
| DeepSWE | 0.9% | estimated ± 6.4 pp, low confidence |
| HLE w/o tools | 5.2% | measured |
| PostTrainBench v1.1 | 8.5% | estimated ± 6.1 pp, low confidence |
| Toolathlon-Verified | 0.0% | estimated ± 2.4 pp, medium confidence |
| Vibe Code Bench | 10.0% | estimated ± 10.3 pp, low confidence |
| CharXiv | 73.9% | estimated ± 3.3 pp, high confidence |
| CharXiv w/o tools | 83.7% | estimated ± 1.7 pp, low confidence |
| C-Eval | 86.6% | estimated ± 0.9 pp, low confidence |
| HMMT Feb 2025 | 63.5% | estimated ± 1.9 pp, low confidence |
| HMMT Feb 2026 | 57.5% | estimated ± 4.0 pp, high confidence |
| MathVision | 79.7% | measured |
| MCP Atlas | 50.2% | estimated ± 11.1 pp, low confidence |
| MCP-Tasks | 100.0% | estimated ± 8.2 pp, low confidence |
| MMAnswerBench | 57.2% | estimated ± 1.3 pp, medium confidence |
| MMLU-Redux | 88.7% | estimated ± 0.7 pp, high confidence |
| MMMU-Pro | 69.1% | measured |
| NL2Repo | 38.9% | estimated ± 3.1 pp, low confidence |
| QwenClawBench | 55.9% | estimated ± 6.0 pp, low confidence |
| ScreenSpot Pro | 55.5% | estimated ± 4.5 pp, medium confidence |
| SuperGPQA | 57.4% | estimated ± 7.5 pp, medium confidence |
| Toolathlon | 26.9% | estimated ± 6.3 pp, low confidence |
| V* | 80.0% | estimated ± 5.1 pp, medium confidence |
| VideoMMMU | 84.1% | estimated ± 1.0 pp, medium confidence |
| ERQA | 54.9% | estimated ± 2.4 pp, low confidence |
| HealthBench Hard | 0.0% | estimated ± 5.1 pp, low confidence |
| MedXpertQA (MM) | 48.7% | measured |
| MedXpertQA (Text) | 31.6% | estimated ± 4.5 pp, low confidence |
| MMLU-Pro (Arcee) | 82.8% | estimated ± 3.4 pp, medium confidence |
| React Native Evals | 71.9% | estimated ± 3.5 pp, high confidence |
| ResearchClawBench | 0.0% | estimated ± 2.2 pp, low confidence |
| SWE-Rebench | 0.0% | estimated ± 8.9 pp, low confidence |
| ARC-AGI-3 | 0.0% | estimated ± 2.1 pp, medium confidence |
| BioMysteryBench (human-solvable) | 81.7% | estimated ± 0.6 pp, low confidence |
| DRACO | 61.5% | estimated ± 3.0 pp, low confidence |
| FrontierCode 1.1 Extended | 29.8% | estimated ± 3.4 pp, low confidence |
| HealthBench (raw) | 22.6% | estimated ± 6.4 pp, low confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.7 pp, low confidence |
| HealthBench Professional | 29.3% | estimated ± 4.8 pp, low confidence |
| HealthBench Professional (raw) | 35.2% | estimated ± 5.0 pp, low confidence |
| HLE-Verified | 15.3% | estimated ± 2.0 pp, low confidence |
| LABBench2 | 56.9% | estimated ± 1.8 pp, low confidence |
| AIME 2025 | 74.2% | estimated ± 1.5 pp, medium confidence |
| MMMU | 78.9% | estimated ± 0.9 pp, high confidence |
| LiveCodeBench | 53.8% | estimated ± 8.9 pp, low confidence |
| OpenHarmony Bench | 38.9% | estimated ± 2.8 pp, medium confidence |
| BabyVision | 19.9% | estimated ± 4.3 pp, low confidence |
| MMVU | 72.1% | estimated ± 2.9 pp, low confidence |
| SimpleVQA | 61.7% | estimated ± 8.2 pp, medium confidence |
| ZeroBench | 8.8% | estimated ± 4.7 pp, low confidence |
| MMMLU | 83.4% | measured |
| MMLU | 82.6% | estimated ± 1.1 pp, medium confidence |
| MMMU-Pro w/ Python | 71.5% | estimated ± 0.6 pp, high confidence |
| MATH-500 | 95.2% | estimated ± 3.1 pp, medium confidence |
| MathVision w/ Python | 89.1% | estimated ± 0.5 pp, low confidence |
| RealWorldQA | 0.0% | estimated ± 0.8 pp, low confidence |
| Video-MME (with subtitle) | 84.2% | estimated ± 1.8 pp, low confidence |
| OmniDocBench 1.5 | 0.1% | estimated ± 5.9 pp, low confidence |