Microsoft · OpenRouter API
Microsoft-Decision-1 by Microsoft scores 82.0 on the System One Index (4th of 24). Calibration error 2.8%, median latency 240 ms.
| Dataset | Accuracy | Rank | Best on the board | Calibration error |
|---|---|---|---|---|
| AG News English | 89.0 | 11th of 24 | 92.5 | 6.7% |
| SST-2 English | 95.5 | 4th of 24 | 97.0 | 2.1% |
| SST-5 English | 62.5 | 1st of 24 | 62.5 | 11.4% |
| DAIR Emotion English | 57.0 | 12th of 24 | 91.5 | 16.4% |
| Banking77 English | 87.0 | 10th of 24 | 98.5 | 3.8% |
| BoolQ English | 88.5 | 11th of 24 | 92.5 | 6.7% |
| XNLI English English | 89.3 | 6th of 24 | 93.3 | 2.6% |
| Typed decisions English | 79.0 | 1st of 24 | 79.0 | 2.6% |
| XNLI French French | 82.0 | 7th of 24 | 90.7 | 11.0% |
| XNLI Arabic Arabic | 72.0 | 13th of 26 | 89.3 | 12.5% |
| MASSIVE scenario Arabic | 83.5 | 4th of 26 | 89.5 | 7.0% |
| MASSIVE intent Arabic | 89.0 | 6th of 26 | 91.5 | 3.5% |
| OSACT4 offensive Arabic | 88.5 | 10th of 26 | 95.5 | 4.5% |
| Tweet sentiment Arabic | 68.5 | 8th of 26 | 80.0 | 14.8% |
| MIRACL rerank Arabic | 90.0 | 2nd of 26 | 90.5 | 4.7% |
| MIRACL pairwise Arabic | 91.0 | 1st of 26 | 91.0 | 1.6% |
microsoft/microsoft-decision-1Ranks count the models charted on the leaderboard. Every model gets the same cases and the same request, under the benchmark rules.