← All comparisons
Llama 3.2 Instruct 11B (Vision) vs Claude 4.5 Haiku (Non-reasoning)
Meta vs Anthropic — side-by-side benchmark comparison
| Llama 3.2 Instruct 11B (Vision) | Claude 4.5 Haiku (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 8.7 | 31.0 |
| Coding Index | 4.2 | 29.6 |
| Math Index | 1.7 | 39.0 |
| Output speed (tok/s) | 77.0 | 105.8 |
| Blended price ($/1M) | $0.24 | $2.19 |
| Time to first token (s) | 0.49s | 0.65s |
| aime | 9.3% | — |
| aime 25 | 1.7% | 39.0% |
| artificial analysis coding index | 4.20 | 29.60 |
| artificial analysis intelligence index | 8.70 | 31.00 |
| artificial analysis math index | 1.70 | 39.00 |
| gpqa | 22.1% | 64.6% |
| hle | 5.2% | 4.3% |
| ifbench | 30.4% | 42.0% |
| lcr | 11.7% | 43.7% |
| livecodebench | 11.0% | 51.1% |
| math 500 | 51.6% | — |
| mmlu pro | 46.4% | 80.0% |
| scicode | 11.2% | 34.4% |
| tau2 | 14.6% | 32.5% |
| terminalbench hard | 0.8% | 27.3% |
Benchmark data from Artificial Analysis.