Head to head
Decider V1.2 scores 84.6 and Span-02 79.5 on the System One Index. Decider V1.2 is more accurate on 11 of 16 text datasets, Span-02 on 3, and 2 are tied. Same cases, same request, same scorer.
| Decider V1.2 | Span-02 | |
|---|---|---|
| System One Index | 84.6 | 79.5 |
| English | 82.0 | 77.9 |
| Arabic | 86.9 | 82.3 |
| Vision Index | 83.8 | 86.1 |
| Calibration error | 0.7% | 10.5% |
| Median latency | 178 ms | 204 ms |
| Cost per 1,000 calls | $0.007 | $0.005 |
| Decider V1.2 | Span-02 | |
|---|---|---|
| AG News English | 91.5 | 86.5 |
| SST-2 English | 95.5 | 95.5 |
| SST-5 English | 60.0 | 57.5 |
| DAIR Emotion English | 60.0 | 56.0 |
| Banking77 English | 88.0 | 82.0 |
| BoolQ English | 91.5 | 90.0 |
| XNLI English English | 94.0 | 83.3 |
| Typed decisions English | 75.1 | 72.1 |
| XNLI French French | 90.0 | 73.3 |
| XNLI Arabic Arabic | 90.0 | 66.0 |
| MASSIVE scenario Arabic | 79.0 | 71.0 |
| MASSIVE intent Arabic | 91.0 | 86.0 |
| OSACT4 offensive Arabic | 95.0 | 97.0 |
| Tweet sentiment Arabic | 75.5 | 76.0 |
| MIRACL rerank Arabic | 86.5 | 88.5 |
| MIRACL pairwise Arabic | 91.5 | 91.5 |
| Decider V1.2 | Span-02 | |
|---|---|---|
| SST-2, as an image | 94.5 | 92.5 |
| Arabic sentiment, as an image | 57.0 | 70.0 |
| Oxford-IIIT Pet | 85.5 | 92.5 |
| VQAv2 yes/no | 86.5 | 87.0 |
| CountBench | 95.5 | 88.5 |
Model pages: Decider V1.2 ยท Span-02. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.