Head to head
Span-02 scores 79.5 and Jev 1.13 79.0 on the System One Index. Span-02 is more accurate on 8 of 16 text datasets, Jev 1.13 on 6, and 2 are tied. Same cases, same request, same scorer.
| Span-02 | Jev 1.13 | |
|---|---|---|
| System One Index | 79.5 | 79.0 |
| English | 77.9 | 78.5 |
| Arabic | 82.3 | 79.9 |
| Vision Index | 86.1 | 83.9 |
| Calibration error | 10.5% | 6.1% |
| Median latency | 204 ms | 150 ms |
| Cost per 1,000 calls | $0.005 | $0.028 |
| Span-02 | Jev 1.13 | |
|---|---|---|
| AG News English | 86.5 | 88.5 |
| SST-2 English | 95.5 | 94.5 |
| SST-5 English | 57.5 | 62.0 |
| DAIR Emotion English | 56.0 | 56.0 |
| Banking77 English | 82.0 | 79.0 |
| BoolQ English | 90.0 | 89.5 |
| XNLI English English | 83.3 | 86.0 |
| Typed decisions English | 72.1 | 72.7 |
| XNLI French French | 73.3 | 76.7 |
| XNLI Arabic Arabic | 66.0 | 69.3 |
| MASSIVE scenario Arabic | 71.0 | 63.0 |
| MASSIVE intent Arabic | 86.0 | 84.0 |
| OSACT4 offensive Arabic | 97.0 | 90.0 |
| Tweet sentiment Arabic | 76.0 | 76.0 |
| MIRACL rerank Arabic | 88.5 | 86.5 |
| MIRACL pairwise Arabic | 91.5 | 90.5 |
| Span-02 | Jev 1.13 | |
|---|---|---|
| SST-2, as an image | 92.5 | 94.5 |
| Arabic sentiment, as an image | 70.0 | 59.0 |
| Oxford-IIIT Pet | 92.5 | 83.5 |
| VQAv2 yes/no | 87.0 | 88.0 |
| CountBench | 88.5 | 94.5 |
Model pages: Span-02 ยท Jev 1.13. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.