← All comparisons
DeepSeek R1 Distill Qwen 1.5B vs Qwen3 VL 30B A3B (Reasoning)
DeepSeek vs Alibaba — side-by-side benchmark comparison
| DeepSeek R1 Distill Qwen 1.5B | Qwen3 VL 30B A3B (Reasoning) | |
|---|---|---|
| Intelligence Index | 9.1 | 19.7 |
| Coding Index | — | 13.1 |
| Math Index | 22.0 | 82.3 |
| Output speed (tok/s) | 0.0 | 123.6 |
| Blended price ($/1M) | $0.00 | $0.34 |
| Time to first token (s) | 0.00s | 1.09s |
| aime | 17.7% | — |
| aime 25 | 22.0% | 82.3% |
| artificial analysis coding index | — | 13.10 |
| artificial analysis intelligence index | 9.10 | 19.70 |
| artificial analysis math index | 22.00 | 82.30 |
| gpqa | 9.8% | 72.0% |
| hle | 3.3% | 8.7% |
| ifbench | 13.2% | 45.1% |
| lcr | 0.3% | 40.7% |
| livecodebench | 7.0% | 69.7% |
| math 500 | 68.7% | — |
| mmlu pro | 26.9% | 80.7% |
| scicode | 6.6% | 28.8% |
| tau2 | — | 19.9% |
| terminalbench hard | — | 5.3% |
Benchmark data from Artificial Analysis.