← All comparisons
Qwen3.5 4B (Non-reasoning) vs GPT-5 (medium)
Alibaba vs OpenAI — side-by-side benchmark comparison
| Qwen3.5 4B (Non-reasoning) | GPT-5 (medium) | |
|---|---|---|
| Intelligence Index | 22.6 | 42.0 |
| Coding Index | 13.7 | 38.9 |
| Math Index | — | 91.7 |
| Output speed (tok/s) | 210.0 | 86.4 |
| Blended price ($/1M) | $0.06 | $3.44 |
| Time to first token (s) | 0.23s | 37.15s |
| aime | — | 91.7% |
| aime 25 | — | 91.7% |
| artificial analysis coding index | 13.70 | 38.90 |
| artificial analysis intelligence index | 22.60 | 42.00 |
| artificial analysis math index | — | 91.70 |
| gpqa | 71.2% | 84.2% |
| hle | 7.5% | 23.5% |
| ifbench | 33.3% | 70.6% |
| lcr | 28.3% | 72.8% |
| livecodebench | — | 70.3% |
| math 500 | — | 99.1% |
| mmlu pro | — | 86.7% |
| scicode | 18.3% | 41.1% |
| tau2 | 87.7% | 86.5% |
| terminalbench hard | 11.4% | 37.9% |
Benchmark data from Artificial Analysis.