Head to head
d1 scores 82.7 and Solar Decide 68.8 on the System One Index. d1 is more accurate on 15 of 16 text datasets, Solar Decide on 1. Same cases, same request, same scorer.
| d1 | Solar Decide | |
|---|---|---|
| System One Index | 82.7 | 68.8 |
| English | 81.3 | 64.7 |
| Arabic | 83.9 | 73.7 |
| Vision Index | – | – |
| Calibration error | 2.9% | 20.3% |
| Median latency | 188 ms | 715 ms |
| Cost per 1,000 calls | $0.015 | $0.042 |
| d1 | Solar Decide | |
|---|---|---|
| AG News English | 90.0 | 86.0 |
| SST-2 English | 95.5 | 91.5 |
| SST-5 English | 56.5 | 53.5 |
| DAIR Emotion English | 63.5 | 54.0 |
| Banking77 English | 89.0 | 1.0 |
| BoolQ English | 90.0 | 87.0 |
| XNLI English English | 91.3 | 74.7 |
| Typed decisions English | 74.5 | 69.9 |
| XNLI French French | 84.7 | 67.3 |
| XNLI Arabic Arabic | 80.0 | 58.0 |
| MASSIVE scenario Arabic | 73.0 | 55.5 |
| MASSIVE intent Arabic | 90.0 | 75.0 |
| OSACT4 offensive Arabic | 95.5 | 86.0 |
| Tweet sentiment Arabic | 67.5 | 71.5 |
| MIRACL rerank Arabic | 90.5 | 84.5 |
| MIRACL pairwise Arabic | 91.0 | 85.5 |
Model pages: d1 · Solar Decide. Latency for API models is the median per call through OpenRouter, timed from one machine.