← All comparisons
Grok 4.20 0309 (Non-reasoning) vs Qwen3 235B A22B 2507 (Reasoning)
xAI vs Alibaba — side-by-side benchmark comparison
| Grok 4.20 0309 (Non-reasoning) | Qwen3 235B A22B 2507 (Reasoning) | |
|---|---|---|
| Intelligence Index | 29.7 | 29.5 |
| Coding Index | 25.4 | 23.2 |
| Math Index | — | 91.0 |
| Output speed (tok/s) | 202.6 | 62.5 |
| Blended price ($/1M) | $3.00 | $0.84 |
| Time to first token (s) | 0.50s | 1.21s |
| aime | — | 94.0% |
| aime 25 | — | 91.0% |
| artificial analysis coding index | 25.40 | 23.20 |
| artificial analysis intelligence index | 29.70 | 29.50 |
| artificial analysis math index | — | 91.00 |
| gpqa | 78.5% | 79.0% |
| hle | 22.5% | 15.0% |
| ifbench | 47.8% | 51.2% |
| lcr | 18.0% | 67.0% |
| livecodebench | — | 78.8% |
| math 500 | — | 98.4% |
| mmlu pro | — | 84.3% |
| scicode | 32.2% | 42.4% |
| tau2 | 69.6% | 53.2% |
| terminalbench hard | 22.0% | 13.6% |
Benchmark data from Artificial Analysis.