← All comparisons
Qwen3.6 27B (Non-reasoning) vs Qwen3 VL 8B Instruct
Alibaba vs Alibaba — side-by-side benchmark comparison
| Qwen3.6 27B (Non-reasoning) | Qwen3 VL 8B Instruct | |
|---|---|---|
| Intelligence Index | 37.1 | 14.3 |
| Coding Index | 26.6 | 7.3 |
| Math Index | — | 27.3 |
| Output speed (tok/s) | 60.7 | 143.8 |
| Blended price ($/1M) | $1.35 | $0.31 |
| Time to first token (s) | 1.55s | 0.93s |
| aime | — | — |
| aime 25 | — | 27.3% |
| artificial analysis coding index | 26.60 | 7.30 |
| artificial analysis intelligence index | 37.10 | 14.30 |
| artificial analysis math index | — | 27.30 |
| gpqa | 82.9% | 42.7% |
| hle | 13.6% | 2.9% |
| ifbench | 45.7% | 32.3% |
| lcr | 55.0% | 15.3% |
| livecodebench | — | 33.2% |
| math 500 | — | — |
| mmlu pro | — | 68.6% |
| scicode | 37.3% | 17.4% |
| tau2 | 93.6% | 29.2% |
| terminalbench hard | 21.2% | 2.3% |
Benchmark data from Artificial Analysis.