Head to head
Microsoft-Decision-1 scores 82.0 and d3 edge 65.9 on the System One Index. Microsoft-Decision-1 is more accurate on 14 of 16 text datasets, d3 edge on 1, and 1 are tied. Same cases, same request, same scorer.
| Microsoft-Decision-1 | d3 edge | |
|---|---|---|
| System One Index | 82.0 | 65.9 |
| English | 81.0 | 69.1 |
| Arabic | 83.2 | 62.2 |
| Vision Index | – | – |
| Calibration error | 2.8% | 18.7% |
| Median latency | 240 ms | – |
| Cost per 1,000 calls | $0.010 | – |
| Microsoft-Decision-1 | d3 edge | |
|---|---|---|
| AG News English | 89.0 | 80.0 |
| SST-2 English | 95.5 | 80.5 |
| SST-5 English | 62.5 | 39.5 |
| DAIR Emotion English | 57.0 | 57.0 |
| Banking77 English | 87.0 | 90.0 |
| BoolQ English | 88.5 | 79.5 |
| XNLI English English | 89.3 | 84.0 |
| Typed decisions English | 79.0 | 42.3 |
| XNLI French French | 82.0 | 66.0 |
| XNLI Arabic Arabic | 72.0 | 59.3 |
| MASSIVE scenario Arabic | 83.5 | 41.0 |
| MASSIVE intent Arabic | 89.0 | 55.0 |
| OSACT4 offensive Arabic | 88.5 | 72.5 |
| Tweet sentiment Arabic | 68.5 | 61.5 |
| MIRACL rerank Arabic | 90.0 | 70.0 |
| MIRACL pairwise Arabic | 91.0 | 76.0 |
Model pages: Microsoft-Decision-1 · d3 edge. Latency for API models is the median per call through OpenRouter, timed from one machine.