Language Understanding on MMLU 10-subject subset 1.0 (test)
85.4Average AccuracyNS-12
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| NS-12Model=Qwen3-80B-A3B2026.05 | 85.4 | 80 | 80.7 | 94.1 | 81 | 89.1 | 97.2 | 68 | 86 | 77 | 86.1 | |
| TritonModel=Qwen3-80B-A3B2026.05 | 85.3 | 80 | 80.7 | 93.4 | 81 | 89.4 | 97.2 | 67 | 85 | 78 | 86.1 | |
| VCSModel=Qwen3-80B-A3B2026.05 | 85.3 | 80 | 81.5 | 93.4 | 81 | 89.4 | 97.2 | 67 | 84 | 77 | 86.7 | |
| MXRModel=Qwen3-80B-A3B2026.05 | 85 | 80 | 80 | 93.4 | 80 | 89.4 | 96.5 | 67 | 85 | 76 | 86.1 | |
| TritonModel=Qwen3-35B-A3B2026.05 | 76.1 | 48 | 83 | 73.7 | 85 | 82.3 | 94.4 | 47 | 77 | 68 | 80.3 | |
| NS-12Model=Qwen3-35B-A3B2026.05 | 75.7 | 46 | 83.7 | 74.3 | 84 | 82.6 | 93.8 | 45 | 79 | 68 | 77.5 | |
| VCSModel=Qwen3-35B-A3B2026.05 | 75.2 | 46 | 83 | 73.7 | 83 | 81.1 | 93.1 | 46 | 76 | 70 | 78 | |
| MXRModel=Qwen3-35B-A3B2026.05 | 75.2 | 50 | 83 | 71.7 | 83 | 81.9 | 95.1 | 44 | 75 | 68 | 77.5 | |
| MXRModel=Qwen3.5-0.8B2026.05 | 49.7 | 33 | 45.2 | 50.7 | 54 | 60 | 59 | 49 | 44 | 31 | 50.3 | |
| NS-12Model=Qwen3.5-0.8B2026.05 | 49.5 | 33 | 46.7 | 49.3 | 53 | 58.9 | 60.4 | 49 | 43 | 33 | 49.1 | |
| VCSModel=Qwen3.5-0.8B2026.05 | 49.3 | 33 | 46.7 | 50.7 | 50 | 58.1 | 60.4 | 49 | 42 | 33 | 50.3 | |
| TritonModel=Qwen3.5-0.8B2026.05 | 49.2 | 32 | 46.7 | 50 | 51 | 58.1 | 61.8 | 47 | 42 | 35 | 49.1 |