~/system-one-bench / compare/baseten-elephant-1-vs-nace-ai-drex-max
Head to head
Elephant 1 scores 82.1 and Drex Max 84.1 on the System One Index. Elephant 1 is more accurate on 4 of 16 text datasets, Drex Max on 12. Same cases, same request, same scorer.
| Elephant 1 | Drex Max | |
|---|---|---|
| System One Index | 82.1 | 84.1 |
| English | 81.3 | 82.8 |
| Arabic | 82.0 | 84.5 |
| Vision Index | โ | โ |
| Calibration error | 3.7% | 6.2% |
| Median latency | โ | 277 ms |
| Cost per 1,000 calls | $0.013 | $0.021 |
| Elephant 1 | Drex Max | |
|---|---|---|
| AG News English | 89.0 | 90.5 |
| SST-2 English | 94.5 | 96.0 |
| SST-5 English | 58.5 | 59.0 |
| DAIR Emotion English | 54.5 | 65.0 |
| Banking77 English | 94.5 | 90.5 |
| BoolQ English | 91.0 | 93.0 |
| XNLI English English | 94.0 | 95.3 |
| Typed decisions English | 74.2 | 73.0 |
| XNLI French French | 90.0 | 91.3 |
| XNLI Arabic Arabic | 79.3 | 89.3 |
| MASSIVE scenario Arabic | 69.0 | 73.0 |
| MASSIVE intent Arabic | 89.5 | 88.0 |
| OSACT4 offensive Arabic | 91.0 | 93.5 |
| Tweet sentiment Arabic | 66.0 | 69.0 |
| MIRACL rerank Arabic | 88.0 | 89.5 |
| MIRACL pairwise Arabic | 91.0 | 89.5 |
Model pages: Elephant 1 ยท Drex Max. Latency for API models is the median per call through OpenRouter, timed from a cloud server with 8 calls in flight.