← All comparisons
Granite 4.1 3B vs Qwen3 VL 32B Instruct
IBM vs Alibaba — side-by-side benchmark comparison
| Granite 4.1 3B | Qwen3 VL 32B Instruct | |
|---|---|---|
| Intelligence Index | 8.5 | 17.2 |
| Coding Index | 5.5 | 15.6 |
| Math Index | — | 68.3 |
| Output speed (tok/s) | 0.0 | 76.1 |
| Blended price ($/1M) | $0.00 | $1.23 |
| Time to first token (s) | 0.00s | 1.30s |
| aime | — | — |
| aime 25 | — | 68.3% |
| artificial analysis coding index | 5.50 | 15.60 |
| artificial analysis intelligence index | 8.50 | 17.20 |
| artificial analysis math index | — | 68.30 |
| gpqa | 31.4% | 67.1% |
| hle | 3.4% | 6.3% |
| ifbench | 33.7% | 39.2% |
| lcr | 3.0% | 31.3% |
| livecodebench | — | 51.4% |
| math 500 | — | — |
| mmlu pro | — | 79.1% |
| scicode | 11.9% | 30.1% |
| tau2 | 19.6% | 29.2% |
| terminalbench hard | 2.3% | 8.3% |
Benchmark data from Artificial Analysis.