← All comparisons
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) vs Qwen3 VL 30B A3B (Reasoning)
Anthropic vs Alibaba — side-by-side benchmark comparison
| Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | Qwen3 VL 30B A3B (Reasoning) | |
|---|---|---|
| Intelligence Index | 51.7 | 19.7 |
| Coding Index | 50.9 | 13.1 |
| Math Index | — | 82.3 |
| Output speed (tok/s) | 68.2 | 123.6 |
| Blended price ($/1M) | $6.56 | $0.34 |
| Time to first token (s) | 55.35s | 1.09s |
| aime | — | — |
| aime 25 | — | 82.3% |
| artificial analysis coding index | 50.90 | 13.10 |
| artificial analysis intelligence index | 51.70 | 19.70 |
| artificial analysis math index | — | 82.30 |
| gpqa | 87.5% | 72.0% |
| hle | 30.0% | 8.7% |
| ifbench | 56.6% | 45.1% |
| lcr | 70.7% | 40.7% |
| livecodebench | — | 69.7% |
| math 500 | — | — |
| mmlu pro | — | 80.7% |
| scicode | 46.8% | 28.8% |
| tau2 | 75.7% | 19.9% |
| terminalbench hard | 53.0% | 5.3% |
Benchmark data from Artificial Analysis.