Head to head
Laya Router scores 60.4 and Mercury Decide 82.4 on the System One Index. Laya Router is more accurate on 1 of 16 text datasets, Mercury Decide on 15. Same cases, same request, same scorer.
| Laya Router | Mercury Decide | |
|---|---|---|
| System One Index | 60.4 | 82.4 |
| English | 64.7 | 83.7 |
| Arabic | 53.2 | 80.8 |
| Vision Index | – | 79.4 |
| Calibration error | 20.0% | 13.1% |
| Median latency | – | 276 ms |
| Cost per 1,000 calls | – | Free |
| Laya Router | Mercury Decide | |
|---|---|---|
| AG News English | 92.5 | 92.0 |
| SST-2 English | 88.5 | 95.5 |
| SST-5 English | 34.0 | 52.0 |
| DAIR Emotion English | 53.5 | 91.5 |
| Banking77 English | 44.0 | 86.0 |
| BoolQ English | 81.5 | 90.5 |
| XNLI English English | 87.3 | 91.3 |
| Typed decisions English | 36.4 | 71.1 |
| XNLI French French | 76.0 | 82.7 |
| XNLI Arabic Arabic | 68.0 | 76.0 |
| MASSIVE scenario Arabic | 39.0 | 72.0 |
| MASSIVE intent Arabic | 32.5 | 85.0 |
| OSACT4 offensive Arabic | 82.5 | 92.0 |
| Tweet sentiment Arabic | 37.0 | 62.5 |
| MIRACL rerank Arabic | 31.5 | 88.0 |
| MIRACL pairwise Arabic | 82.0 | 90.0 |
| Laya Router | Mercury Decide | |
|---|---|---|
| SST-2, as an image | – | 93.0 |
| Arabic sentiment, as an image | – | 39.0 |
| Oxford-IIIT Pet | – | 84.5 |
| VQAv2 yes/no | – | 87.0 |
| CountBench | – | 93.5 |
Model pages: Laya Router · Mercury Decide. Latency for API models is the median per call through OpenRouter, timed from one machine.