Head to head
Clef scores 81.6 and Solar Decide Flash 63.6 on the System One Index. Clef is more accurate on 15 of 16 text datasets, Solar Decide Flash on 1. Same cases, same request, same scorer.
| Clef | Solar Decide Flash | |
|---|---|---|
| System One Index | 81.6 | 63.6 |
| English | 80.6 | 61.3 |
| Arabic | 83.0 | 66.5 |
| Vision Index | 78.1 | – |
| Calibration error | 2.2% | 14.4% |
| Median latency | 310 ms | 586 ms |
| Cost per 1,000 calls | $0.116 | $0.042 |
| Clef | Solar Decide Flash | |
|---|---|---|
| AG News English | 90.5 | 86.0 |
| SST-2 English | 97.0 | 90.5 |
| SST-5 English | 51.5 | 44.5 |
| DAIR Emotion English | 58.5 | 54.5 |
| Banking77 English | 95.5 | 1.0 |
| BoolQ English | 91.0 | 86.0 |
| XNLI English English | 88.7 | 65.3 |
| Typed decisions English | 72.2 | 62.5 |
| XNLI French French | 80.0 | 61.3 |
| XNLI Arabic Arabic | 78.0 | 52.0 |
| MASSIVE scenario Arabic | 88.0 | 46.0 |
| MASSIVE intent Arabic | 89.5 | 63.0 |
| OSACT4 offensive Arabic | 89.5 | 84.5 |
| Tweet sentiment Arabic | 58.0 | 65.5 |
| MIRACL rerank Arabic | 89.0 | 71.5 |
| MIRACL pairwise Arabic | 89.0 | 83.0 |
| Clef | Solar Decide Flash | |
|---|---|---|
| SST-2, as an image | 95.5 | – |
| Arabic sentiment, as an image | 35.0 | – |
| Oxford-IIIT Pet | 82.5 | – |
| VQAv2 yes/no | 87.5 | – |
| CountBench | 90.0 | – |
Model pages: Clef · Solar Decide Flash. Latency for API models is the median per call through OpenRouter, timed from one machine.