← All comparisons
Phi-4 Mini Instruct vs Claude 4.1 Opus (Reasoning)
Microsoft vs Anthropic — side-by-side benchmark comparison
| Phi-4 Mini Instruct | Claude 4.1 Opus (Reasoning) | |
|---|---|---|
| Intelligence Index | 8.4 | 42.0 |
| Coding Index | 3.6 | 36.5 |
| Math Index | 6.7 | 80.3 |
| Output speed (tok/s) | 0.0 | 44.5 |
| Blended price ($/1M) | $0.00 | $32.81 |
| Time to first token (s) | 0.00s | 8.55s |
| aime | 3.0% | — |
| aime 25 | 6.7% | 80.3% |
| artificial analysis coding index | 3.60 | 36.50 |
| artificial analysis intelligence index | 8.40 | 42.00 |
| artificial analysis math index | 6.70 | 80.30 |
| gpqa | 33.1% | 80.9% |
| hle | 4.2% | 11.9% |
| ifbench | 21.1% | 55.4% |
| lcr | 13.7% | 66.3% |
| livecodebench | 12.6% | 65.4% |
| math 500 | 69.6% | — |
| mmlu pro | 46.5% | 88.0% |
| scicode | 10.8% | 40.9% |
| tau2 | 8.2% | 71.4% |
| terminalbench hard | 0.0% | 34.3% |
Benchmark data from Artificial Analysis.