← All comparisons
Llama 3.2 Instruct 3B vs Claude 4.1 Opus (Reasoning)
Meta vs Anthropic — side-by-side benchmark comparison
| Llama 3.2 Instruct 3B | Claude 4.1 Opus (Reasoning) | |
|---|---|---|
| Intelligence Index | 9.7 | 42.0 |
| Coding Index | — | 36.5 |
| Math Index | 3.3 | 80.3 |
| Output speed (tok/s) | 52.7 | 44.5 |
| Blended price ($/1M) | $0.15 | $32.81 |
| Time to first token (s) | 0.65s | 8.55s |
| aime | 6.7% | — |
| aime 25 | 3.3% | 80.3% |
| artificial analysis coding index | — | 36.50 |
| artificial analysis intelligence index | 9.70 | 42.00 |
| artificial analysis math index | 3.30 | 80.30 |
| gpqa | 25.5% | 80.9% |
| hle | 5.2% | 11.9% |
| ifbench | 26.2% | 55.4% |
| lcr | 2.0% | 66.3% |
| livecodebench | 8.3% | 65.4% |
| math 500 | 48.9% | — |
| mmlu pro | 34.7% | 88.0% |
| scicode | 5.2% | 40.9% |
| tau2 | 21.1% | 71.4% |
| terminalbench hard | — | 34.3% |
Benchmark data from Artificial Analysis.