← All comparisons
Qwen3.6 27B (Non-reasoning) vs DeepSeek-V2.5
Alibaba vs DeepSeek — side-by-side benchmark comparison
| Qwen3.6 27B (Non-reasoning) | DeepSeek-V2.5 | |
|---|---|---|
| Intelligence Index | 37.1 | 12.3 |
| Coding Index | 26.6 | — |
| Math Index | — | — |
| Output speed (tok/s) | 60.7 | 0.0 |
| Blended price ($/1M) | $1.35 | $0.00 |
| Time to first token (s) | 1.55s | 0.00s |
| aime | — | — |
| aime 25 | — | — |
| artificial analysis coding index | 26.60 | — |
| artificial analysis intelligence index | 37.10 | 12.30 |
| artificial analysis math index | — | — |
| gpqa | 82.9% | — |
| hle | 13.6% | — |
| ifbench | 45.7% | — |
| lcr | 55.0% | — |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | — |
| scicode | 37.3% | — |
| tau2 | 93.6% | — |
| terminalbench hard | 21.2% | — |
Benchmark data from Artificial Analysis.