← All comparisons
o3 vs Claude 3.5 Sonnet (Oct '24)
OpenAI vs Anthropic — side-by-side benchmark comparison
| o3 | Claude 3.5 Sonnet (Oct '24) | |
|---|---|---|
| Intelligence Index | 38.4 | 15.9 |
| Coding Index | 38.4 | 30.2 |
| Math Index | 88.3 | — |
| Output speed (tok/s) | 126.8 | 0.0 |
| Blended price ($/1M) | $3.50 | $6.56 |
| Time to first token (s) | 7.82s | 0.00s |
| aime | 90.3% | 15.7% |
| aime 25 | 88.3% | — |
| artificial analysis coding index | 38.40 | 30.20 |
| artificial analysis intelligence index | 38.40 | 15.90 |
| artificial analysis math index | 88.30 | — |
| gpqa | 82.7% | 59.9% |
| hle | 20.0% | 3.9% |
| ifbench | 71.4% | — |
| lcr | 69.3% | — |
| livecodebench | 80.8% | 38.1% |
| math 500 | 99.2% | 77.1% |
| mmlu pro | 85.3% | 77.2% |
| scicode | 41.0% | 36.6% |
| tau2 | 80.7% | — |
| terminalbench hard | 37.1% | — |
Benchmark data from Artificial Analysis.