← All comparisons
Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) vs Qwen3 VL 32B (Reasoning)
Google vs Alibaba — side-by-side benchmark comparison
| Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) | Qwen3 VL 32B (Reasoning) | |
|---|---|---|
| Intelligence Index | 21.6 | 24.7 |
| Coding Index | 18.2 | 14.5 |
| Math Index | 68.7 | 84.7 |
| Output speed (tok/s) | 0.0 | 96.3 |
| Blended price ($/1M) | $0.17 | $2.63 |
| Time to first token (s) | 0.00s | 1.12s |
| aime | — | — |
| aime 25 | 68.7% | 84.7% |
| artificial analysis coding index | 18.20 | 14.50 |
| artificial analysis intelligence index | 21.60 | 24.70 |
| artificial analysis math index | 68.70 | 84.70 |
| gpqa | 70.9% | 73.3% |
| hle | 6.6% | 9.6% |
| ifbench | 52.6% | 59.4% |
| lcr | 59.0% | 55.3% |
| livecodebench | 68.8% | 73.8% |
| math 500 | — | — |
| mmlu pro | 80.8% | 81.8% |
| scicode | 28.7% | 28.5% |
| tau2 | 30.7% | 45.6% |
| terminalbench hard | 12.9% | 7.6% |
Benchmark data from Artificial Analysis.