← All comparisons
Phi-4 vs Claude 4.1 Opus (Non-reasoning)
Microsoft vs Anthropic — side-by-side benchmark comparison
| Phi-4 | Claude 4.1 Opus (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 10.4 | 36.0 |
| Coding Index | 11.2 | — |
| Math Index | 18.0 | — |
| Output speed (tok/s) | 41.1 | 44.7 |
| Blended price ($/1M) | $0.22 | $32.81 |
| Time to first token (s) | 0.50s | 1.63s |
| aime | 14.3% | — |
| aime 25 | 18.0% | — |
| artificial analysis coding index | 11.20 | — |
| artificial analysis intelligence index | 10.40 | 36.00 |
| artificial analysis math index | 18.00 | — |
| gpqa | 57.5% | — |
| hle | 4.1% | — |
| ifbench | 23.5% | — |
| lcr | 0.0% | — |
| livecodebench | 23.1% | — |
| math 500 | 81.0% | — |
| mmlu pro | 71.4% | — |
| scicode | 26.0% | — |
| tau2 | 0.0% | — |
| terminalbench hard | 3.8% | — |
Benchmark data from Artificial Analysis.