← All comparisons
Claude Opus 4.7 (Non-reasoning, High Effort) vs Qwen3 4B 2507 Instruct
Anthropic vs Alibaba — side-by-side benchmark comparison
| Claude Opus 4.7 (Non-reasoning, High Effort) | Qwen3 4B 2507 Instruct | |
|---|---|---|
| Intelligence Index | 51.8 | 12.9 |
| Coding Index | 53.1 | 9.0 |
| Math Index | — | 52.3 |
| Output speed (tok/s) | 47.8 | 0.0 |
| Blended price ($/1M) | $10.94 | $0.00 |
| Time to first token (s) | 1.04s | 0.00s |
| aime | — | — |
| aime 25 | — | 52.3% |
| artificial analysis coding index | 53.10 | 9.00 |
| artificial analysis intelligence index | 51.80 | 12.90 |
| artificial analysis math index | — | 52.30 |
| gpqa | 88.5% | 51.7% |
| hle | 31.2% | 4.7% |
| ifbench | 43.6% | 33.5% |
| lcr | 67.0% | 7.3% |
| livecodebench | — | 37.7% |
| math 500 | — | — |
| mmlu pro | — | 67.2% |
| scicode | 50.1% | 18.1% |
| tau2 | 74.0% | 26.6% |
| terminalbench hard | 54.5% | 4.5% |
Benchmark data from Artificial Analysis.