← All comparisons
GPT-5 mini (medium) vs Claude 3.5 Sonnet (June '24)
OpenAI vs Anthropic — side-by-side benchmark comparison
| GPT-5 mini (medium) | Claude 3.5 Sonnet (June '24) | |
|---|---|---|
| Intelligence Index | 38.9 | 14.2 |
| Coding Index | 32.8 | 26.0 |
| Math Index | 85.0 | — |
| Output speed (tok/s) | 92.4 | 0.0 |
| Blended price ($/1M) | $0.69 | $6.56 |
| Time to first token (s) | 16.52s | 0.00s |
| aime | — | 9.7% |
| aime 25 | 85.0% | — |
| artificial analysis coding index | 32.80 | 26.00 |
| artificial analysis intelligence index | 38.90 | 14.20 |
| artificial analysis math index | 85.00 | — |
| gpqa | 80.3% | 56.0% |
| hle | 14.6% | 3.7% |
| ifbench | 71.2% | — |
| lcr | 66.0% | — |
| livecodebench | 69.2% | — |
| math 500 | — | 69.5% |
| mmlu pro | 82.8% | 75.1% |
| scicode | 41.0% | 31.6% |
| tau2 | 71.1% | — |
| terminalbench hard | 28.8% | — |
Benchmark data from Artificial Analysis.