← All comparisons
Claude 4.5 Haiku (Non-reasoning) vs Llama 3.1 Nemotron Instruct 70B
Anthropic vs NVIDIA — side-by-side benchmark comparison
| Claude 4.5 Haiku (Non-reasoning) | Llama 3.1 Nemotron Instruct 70B | |
|---|---|---|
| Intelligence Index | 31.0 | 13.4 |
| Coding Index | 29.6 | 10.8 |
| Math Index | 39.0 | 11.0 |
| Output speed (tok/s) | 105.8 | 292.9 |
| Blended price ($/1M) | $2.19 | $1.20 |
| Time to first token (s) | 0.65s | 0.24s |
| aime | — | 24.7% |
| aime 25 | 39.0% | 11.0% |
| artificial analysis coding index | 29.60 | 10.80 |
| artificial analysis intelligence index | 31.00 | 13.40 |
| artificial analysis math index | 39.00 | 11.00 |
| gpqa | 64.6% | 46.5% |
| hle | 4.3% | 4.6% |
| ifbench | 42.0% | 30.7% |
| lcr | 43.7% | 7.0% |
| livecodebench | 51.1% | 16.9% |
| math 500 | — | 73.3% |
| mmlu pro | 80.0% | 69.0% |
| scicode | 34.4% | 23.3% |
| tau2 | 32.5% | 23.1% |
| terminalbench hard | 27.3% | 4.5% |
Benchmark data from Artificial Analysis.