← All comparisons
Claude 3 Haiku vs Qwen3 30B A3B 2507 (Reasoning)
Anthropic vs Alibaba — side-by-side benchmark comparison
| Claude 3 Haiku | Qwen3 30B A3B 2507 (Reasoning) | |
|---|---|---|
| Intelligence Index | 12.3 | 22.4 |
| Coding Index | 6.7 | 14.6 |
| Math Index | — | 56.3 |
| Output speed (tok/s) | 0.0 | 155.3 |
| Blended price ($/1M) | $0.50 | $0.67 |
| Time to first token (s) | 0.00s | 1.02s |
| aime | 1.0% | 90.7% |
| aime 25 | — | 56.3% |
| artificial analysis coding index | 6.70 | 14.60 |
| artificial analysis intelligence index | 12.30 | 22.40 |
| artificial analysis math index | — | 56.30 |
| gpqa | 37.4% | 70.7% |
| hle | 3.9% | 9.8% |
| ifbench | 36.1% | 50.7% |
| lcr | 21.0% | 59.0% |
| livecodebench | 15.4% | 70.7% |
| math 500 | 39.4% | 97.6% |
| mmlu pro | — | 80.5% |
| scicode | 18.6% | 33.3% |
| tau2 | 21.1% | 28.1% |
| terminalbench hard | 0.8% | 5.3% |
Benchmark data from Artificial Analysis.