← All comparisons
Llama 3.1 Instruct 8B vs Claude 4.1 Opus (Reasoning)
Meta vs Anthropic — side-by-side benchmark comparison
| Llama 3.1 Instruct 8B | Claude 4.1 Opus (Reasoning) | |
|---|---|---|
| Intelligence Index | 11.8 | 42.0 |
| Coding Index | 4.9 | 36.5 |
| Math Index | 4.3 | 80.3 |
| Output speed (tok/s) | 212.7 | 44.5 |
| Blended price ($/1M) | $0.10 | $32.81 |
| Time to first token (s) | 0.45s | 8.55s |
| aime | 7.7% | — |
| aime 25 | 4.3% | 80.3% |
| artificial analysis coding index | 4.90 | 36.50 |
| artificial analysis intelligence index | 11.80 | 42.00 |
| artificial analysis math index | 4.30 | 80.30 |
| gpqa | 25.9% | 80.9% |
| hle | 5.1% | 11.9% |
| ifbench | 28.6% | 55.4% |
| lcr | 15.7% | 66.3% |
| livecodebench | 11.6% | 65.4% |
| math 500 | 51.9% | — |
| mmlu pro | 47.6% | 88.0% |
| scicode | 13.2% | 40.9% |
| tau2 | 16.4% | 71.4% |
| terminalbench hard | 0.8% | 34.3% |
Benchmark data from Artificial Analysis.