← All comparisons
Llama 4 Maverick vs Claude 4.5 Haiku (Non-reasoning)
Meta vs Anthropic — side-by-side benchmark comparison
| Llama 4 Maverick | Claude 4.5 Haiku (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 18.4 | 31.0 |
| Coding Index | 15.6 | 29.6 |
| Math Index | 19.3 | 39.0 |
| Output speed (tok/s) | 119.7 | 105.8 |
| Blended price ($/1M) | $0.47 | $2.19 |
| Time to first token (s) | 0.64s | 0.65s |
| aime | 39.0% | — |
| aime 25 | 19.3% | 39.0% |
| artificial analysis coding index | 15.60 | 29.60 |
| artificial analysis intelligence index | 18.40 | 31.00 |
| artificial analysis math index | 19.30 | 39.00 |
| gpqa | 67.1% | 64.6% |
| hle | 4.8% | 4.3% |
| ifbench | 43.0% | 42.0% |
| lcr | 46.0% | 43.7% |
| livecodebench | 39.7% | 51.1% |
| math 500 | 88.9% | — |
| mmlu pro | 80.9% | 80.0% |
| scicode | 33.1% | 34.4% |
| tau2 | 17.8% | 32.5% |
| terminalbench hard | 6.8% | 27.3% |
Benchmark data from Artificial Analysis.