~/system-one-bench / compare/baseten-joey-1-vs-cloudflare-clef-omni
Head to head
Joey 1 scores 73.3 and Clef Omni 79.7 on the System One Index. Joey 1 is more accurate on 2 of 16 text datasets, Clef Omni on 14. Same cases, same request, same scorer.
| Joey 1 | Clef Omni | |
|---|---|---|
| System One Index | 73.3 | 79.7 |
| English | 72.9 | 78.6 |
| Arabic | 73.5 | 80.5 |
| Vision Index | – | 72.2 |
| Calibration error | 2.1% | 3.4% |
| Median latency | – | 208 ms |
| Cost per 1,000 calls | $0.004 | $0.074 |
| Joey 1 | Clef Omni | |
|---|---|---|
| AG News English | 85.0 | 89.0 |
| SST-2 English | 86.0 | 97.0 |
| SST-5 English | 40.5 | 51.0 |
| DAIR Emotion English | 52.5 | 56.0 |
| Banking77 English | 93.0 | 96.0 |
| BoolQ English | 80.0 | 90.0 |
| XNLI English English | 84.0 | 82.7 |
| Typed decisions English | 62.4 | 67.0 |
| XNLI French French | 74.7 | 82.7 |
| XNLI Arabic Arabic | 64.7 | 73.3 |
| MASSIVE scenario Arabic | 71.5 | 89.0 |
| MASSIVE intent Arabic | 84.5 | 91.5 |
| OSACT4 offensive Arabic | 83.0 | 90.5 |
| Tweet sentiment Arabic | 66.0 | 50.0 |
| MIRACL rerank Arabic | 62.0 | 83.0 |
| MIRACL pairwise Arabic | 83.0 | 86.5 |
| Joey 1 | Clef Omni | |
|---|---|---|
| SST-2, as an image | – | 89.5 |
| Arabic sentiment, as an image | – | 28.5 |
| Oxford-IIIT Pet | – | 88.0 |
| VQAv2 yes/no | – | 87.0 |
| CountBench | – | 68.0 |
Model pages: Joey 1 · Clef Omni. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.