Head to head
Drex v2.0 scores 78.6 and Decider V1.2 84.6 on the System One Index. Drex v2.0 is more accurate on 2 of 16 text datasets, Decider V1.2 on 14. Same cases, same request, same scorer.
| Drex v2.0 | Decider V1.2 | |
|---|---|---|
| System One Index | 78.6 | 84.6 |
| English | 77.8 | 82.0 |
| Arabic | 79.6 | 86.9 |
| Vision Index | – | 83.8 |
| Calibration error | 1.5% | 0.7% |
| Median latency | 294 ms | 178 ms |
| Cost per 1,000 calls | $0.015 | $0.007 |
| Drex v2.0 | Decider V1.2 | |
|---|---|---|
| AG News English | 88.0 | 91.5 |
| SST-2 English | 94.0 | 95.5 |
| SST-5 English | 56.0 | 60.0 |
| DAIR Emotion English | 55.5 | 60.0 |
| Banking77 English | 80.5 | 88.0 |
| BoolQ English | 92.5 | 91.5 |
| XNLI English English | 87.3 | 94.0 |
| Typed decisions English | 68.7 | 75.1 |
| XNLI French French | 77.3 | 90.0 |
| XNLI Arabic Arabic | 71.3 | 90.0 |
| MASSIVE scenario Arabic | 63.5 | 79.0 |
| MASSIVE intent Arabic | 82.5 | 91.0 |
| OSACT4 offensive Arabic | 94.5 | 95.0 |
| Tweet sentiment Arabic | 69.5 | 75.5 |
| MIRACL rerank Arabic | 87.0 | 86.5 |
| MIRACL pairwise Arabic | 89.0 | 91.5 |
| Drex v2.0 | Decider V1.2 | |
|---|---|---|
| SST-2, as an image | – | 94.5 |
| Arabic sentiment, as an image | – | 57.0 |
| Oxford-IIIT Pet | – | 85.5 |
| VQAv2 yes/no | – | 86.5 |
| CountBench | – | 95.5 |
Model pages: Drex v2.0 · Decider V1.2. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.