Head to head
Classify-1 scores 30.6 and d3 edge 65.9 on the System One Index. Classify-1 is more accurate on 2 of 16 text datasets, d3 edge on 14. Same cases, same request, same scorer.
| Classify-1 | d3 edge | |
|---|---|---|
| System One Index | 30.6 | 65.9 |
| English | 30.7 | 69.1 |
| Arabic | 30.6 | 62.2 |
| Vision Index | – | – |
| Calibration error | 3.2% | 18.7% |
| Median latency | 254 ms | – |
| Cost per 1,000 calls | $0.018 | – |
| Classify-1 | d3 edge | |
|---|---|---|
| AG News English | 26.5 | 80.0 |
| SST-2 English | 50.0 | 80.5 |
| SST-5 English | 13.0 | 39.5 |
| DAIR Emotion English | 59.0 | 57.0 |
| Banking77 English | 1.0 | 90.0 |
| BoolQ English | 38.0 | 79.5 |
| XNLI English English | 30.7 | 84.0 |
| Typed decisions English | 27.2 | 42.3 |
| XNLI French French | 30.7 | 66.0 |
| XNLI Arabic Arabic | 30.7 | 59.3 |
| MASSIVE scenario Arabic | 3.0 | 41.0 |
| MASSIVE intent Arabic | 3.5 | 55.0 |
| OSACT4 offensive Arabic | 80.0 | 72.5 |
| Tweet sentiment Arabic | 35.0 | 61.5 |
| MIRACL rerank Arabic | 12.0 | 70.0 |
| MIRACL pairwise Arabic | 50.0 | 76.0 |
Model pages: Classify-1 · d3 edge. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.