← All comparisons
GPT-5.3 Codex (xhigh) vs Qwen3.5 27B (Non-reasoning)
OpenAI vs Alibaba — side-by-side benchmark comparison
| GPT-5.3 Codex (xhigh) | Qwen3.5 27B (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 53.6 | 37.2 |
| Coding Index | 53.1 | 33.4 |
| Math Index | — | — |
| Output speed (tok/s) | 84.7 | 95.3 |
| Blended price ($/1M) | $4.81 | $0.88 |
| Time to first token (s) | 60.27s | 1.40s |
| aime | — | — |
| aime 25 | — | — |
| artificial analysis coding index | 53.10 | 33.40 |
| artificial analysis intelligence index | 53.60 | 37.20 |
| artificial analysis math index | — | — |
| gpqa | 91.5% | 84.2% |
| hle | 39.9% | 13.2% |
| ifbench | 75.4% | 46.9% |
| lcr | 74.0% | 55.7% |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | — |
| scicode | 53.2% | 36.7% |
| tau2 | 86.0% | 87.1% |
| terminalbench hard | 53.0% | 31.8% |
Benchmark data from Artificial Analysis.