← All comparisons
Claude 4.5 Haiku (Reasoning) vs Nova 2.0 Lite (medium)
Anthropic vs Amazon — side-by-side benchmark comparison
| Claude 4.5 Haiku (Reasoning) | Nova 2.0 Lite (medium) | |
|---|---|---|
| Intelligence Index | 37.1 | 29.7 |
| Coding Index | 32.6 | 23.9 |
| Math Index | 83.7 | 88.7 |
| Output speed (tok/s) | 142.2 | 168.5 |
| Blended price ($/1M) | $2.19 | $0.85 |
| Time to first token (s) | 10.48s | 13.49s |
| aime | — | — |
| aime 25 | 83.7% | 88.7% |
| artificial analysis coding index | 32.60 | 23.90 |
| artificial analysis intelligence index | 37.10 | 29.70 |
| artificial analysis math index | 83.70 | 88.70 |
| gpqa | 67.2% | 76.8% |
| hle | 9.7% | 8.6% |
| ifbench | 54.3% | 68.5% |
| lcr | 70.3% | 58.3% |
| livecodebench | 61.5% | 66.3% |
| math 500 | — | — |
| mmlu pro | 76.0% | 81.3% |
| scicode | 43.3% | 36.8% |
| tau2 | 54.7% | 75.7% |
| terminalbench hard | 27.3% | 17.4% |
Benchmark data from Artificial Analysis.