← All comparisons
GLM-4.5-Air vs Qwen3.5 27B (Non-reasoning)
Z AI vs Alibaba — side-by-side benchmark comparison
| GLM-4.5-Air | Qwen3.5 27B (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 23.2 | 37.2 |
| Coding Index | 23.8 | 33.4 |
| Math Index | 80.7 | — |
| Output speed (tok/s) | 86.9 | 95.3 |
| Blended price ($/1M) | $0.37 | $0.88 |
| Time to first token (s) | 1.72s | 1.40s |
| aime | 67.3% | — |
| aime 25 | 80.7% | — |
| artificial analysis coding index | 23.80 | 33.40 |
| artificial analysis intelligence index | 23.20 | 37.20 |
| artificial analysis math index | 80.70 | — |
| gpqa | 73.3% | 84.2% |
| hle | 6.8% | 13.2% |
| ifbench | 37.6% | 46.9% |
| lcr | 43.7% | 55.7% |
| livecodebench | 68.4% | — |
| math 500 | 96.5% | — |
| mmlu pro | 81.5% | — |
| scicode | 30.6% | 36.7% |
| tau2 | 46.5% | 87.1% |
| terminalbench hard | 20.5% | 31.8% |
Benchmark data from Artificial Analysis.