Other · model
Sakana Fugu benchmark scores
As of 2026-11-22, Sakana Fugu (Other) has measured scores on 10 benchmarks and estimated scores on 71 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 91.7% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 53.7% | estimated ± 3.9 pp, high confidence |
| AA-SciCode | 58.7% | estimated ± 3.0 pp, medium confidence |
| Artificial Analysis Intelligence Index | 38.1% | estimated ± 5.6 pp, medium confidence |
| CritPt | 30.3% | estimated ± 4.0 pp, high confidence |
| GPQA Diamond | 95.5% | measured |
| HLE | 54.9% | estimated ± 6.5 pp, medium confidence |
| AA-LCR | 74.1% | estimated ± 3.4 pp, high confidence |
| LiveCodeBench v6 | 92.9% | measured |
| AA-Omniscience Accuracy | 46.2% | estimated ± 7.7 pp, medium confidence |
| SciCode | 60.1% | measured |
| Terminal-Bench 2.1 | 80.2% | measured |
| AA Coding Index | 67.3% | estimated ± 5.7 pp, low confidence |
| AA-MMMU-Pro | 78.3% | estimated ± 4.9 pp, medium confidence |
| SWE-bench Verified | 80.0% | estimated ± 3.6 pp, high confidence |
| SWE-bench Pro | 59.0% | measured |
| MMLU-Pro | 85.8% | estimated ± 2.5 pp, low confidence |
| GPQA | 95.5% | measured |
| MMLU | 88.1% | estimated ± 4.2 pp, medium confidence |
| Vals GPQA Diamond | 94.3% | estimated ± 2.1 pp, medium confidence |
| Vals LiveCodeBench | 88.8% | estimated ± 0.8 pp, low confidence |
| Vals MMLU-Pro | 89.1% | estimated ± 0.9 pp, medium confidence |
| MMMLU | 91.4% | estimated ± 1.1 pp, low confidence |
| MMMU | 87.1% | estimated ± 3.7 pp, low confidence |
| AA MMLU-Pro | 87.4% | estimated ± 1.2 pp, low confidence |
| AA Terminal-Bench 2.1 | 81.7% | estimated ± 2.6 pp, high confidence |
| ARC-AGI-1 | 96.7% | estimated ± 2.5 pp, high confidence |
| ARC-AGI-2 | 87.1% | estimated ± 9.7 pp, medium confidence |
| CursorBench 3.1 | 54.0% | estimated ± 5.6 pp, low confidence |
| CursorBench 3.2 | 56.8% | estimated ± 3.6 pp, high confidence |
| Terminal-Bench 3.0 | 13.6% | estimated ± 9.1 pp, low confidence |
| FrontierCode 1.1 Main | 39.2% | estimated ± 5.0 pp, high confidence |
| FrontierSWE v2 | 13.9% | estimated ± 8.3 pp, medium confidence |
| ParseBench-1 | 62.1% | estimated ± 8.6 pp, low confidence |
| Terminal-Bench 4.1 | 17.3% | estimated ± 10.9 pp, low confidence |
| Vals SWE-bench | 78.0% | estimated ± 5.3 pp, low confidence |
| Terminal-Bench 2.1 (Vals AI) | 67.7% | estimated ± 7.3 pp, medium confidence |
| HealthBench Professional (length-adjusted, Artificial Analysis) | 58.2% | estimated ± 4.4 pp, medium confidence |
| HealthBench Professional (raw, Artificial Analysis) | 66.2% | estimated ± 2.6 pp, medium confidence |
| Bug Hunt Bench | 30.8% | estimated ± 12.1 pp, low confidence |
| CursorBench 4.0 | 34.8% | estimated ± 3.3 pp, medium confidence |
| DeepSWE | 66.4% | estimated ± 8.4 pp, low confidence |
| FrontierCode 2.0 | 20.9% | estimated ± 6.7 pp, medium confidence |
| HLE w/o tools | 47.2% | measured |
| PostTrainBench v1.1 | 25.1% | estimated ± 4.4 pp, high confidence |
| VulcanBench Frontier v4 | 79.1% | estimated ± 8.3 pp, low confidence |
| AA Terminal-Bench 4.0 | 4.3% | estimated ± 11.0 pp, low confidence |
| ARC-AGI-3 | 8.3% | estimated ± 7.2 pp, medium confidence |
| CharXiv | 85.1% | measured |
| CharXiv w/o tools | 79.2% | estimated ± 2.4 pp, high confidence |
| AI2D_TEST | 93.5% | estimated ± 0.9 pp, low confidence |
| C-Eval | 93.7% | estimated ± 1.2 pp, medium confidence |
| CountBench | 97.1% | estimated ± 3.7 pp, low confidence |
| DynaMath | 87.1% | estimated ± 0.8 pp, low confidence |
| ERQA | 68.6% | estimated ± 5.6 pp, medium confidence |
| MathVision | 88.4% | estimated ± 3.3 pp, high confidence |
| MLVU (M-Avg) | 87.3% | estimated ± 1.5 pp, high confidence |
| MMLU-Redux | 95.6% | estimated ± 1.7 pp, medium confidence |
| MMMU-Pro | 79.7% | estimated ± 3.1 pp, high confidence |
| NL2Repo | 43.8% | estimated ± 2.8 pp, medium confidence |
| PostTrainBench v1 | 24.2% | estimated ± 4.6 pp, high confidence |
| RealWorldQA | 85.5% | estimated ± 1.5 pp, high confidence |
| ScreenSpot Pro | 79.3% | estimated ± 6.8 pp, medium confidence |
| SuperGPQA | 82.8% | estimated ± 8.5 pp, low confidence |
| Video-MME (with subtitle) | 88.7% | estimated ± 2.9 pp, high confidence |
| VideoMMMU | 86.2% | estimated ± 1.5 pp, medium confidence |
| We-Math | 84.6% | estimated ± 7.9 pp, low confidence |
| HealthBench Hard | 36.6% | estimated ± 5.1 pp, low confidence |
| LiveCodeBench Pro | 87.8% | measured |
| MedXpertQA (MM) | 77.1% | estimated ± 6.4 pp, low confidence |
| MedXpertQA (Text) | 75.1% | estimated ± 4.5 pp, low confidence |
| MMLU-Pro (Arcee) | 88.7% | estimated ± 3.4 pp, low confidence |
| PostTrainBench | 23.1% | estimated ± 3.7 pp, high confidence |
| FrontierCode 1.1 Extended | 55.7% | estimated ± 3.0 pp, low confidence |
| HealthBench Professional | 55.7% | estimated ± 4.2 pp, medium confidence |
| SimpleQA | 38.6% | estimated ± 13.4 pp, low confidence |
| OpenHarmony Bench v1.0 | 53.3% | estimated ± 4.9 pp, low confidence |
| BabyVision | 67.7% | estimated ± 11.5 pp, low confidence |
| MMVU | 78.5% | estimated ± 3.6 pp, medium confidence |
| MRCRv2 | 86.6% | measured |
| BioBasicsBench | 57.4% | estimated ± 2.2 pp, medium confidence |