← All comparisons
Granite 4.0 H Small vs Qwen3 VL 235B A22B (Reasoning)
IBM vs Alibaba — side-by-side benchmark comparison
| Granite 4.0 H Small | Qwen3 VL 235B A22B (Reasoning) | |
|---|---|---|
| Intelligence Index | 10.8 | 27.6 |
| Coding Index | 8.5 | 20.9 |
| Math Index | 13.7 | 88.3 |
| Output speed (tok/s) | 401.2 | 35.6 |
| Blended price ($/1M) | $0.11 | $2.17 |
| Time to first token (s) | 8.75s | 5.14s |
| aime | — | — |
| aime 25 | 13.7% | 88.3% |
| artificial analysis coding index | 8.50 | 20.90 |
| artificial analysis intelligence index | 10.80 | 27.60 |
| artificial analysis math index | 13.70 | 88.30 |
| gpqa | 41.6% | 77.2% |
| hle | 3.7% | 10.1% |
| ifbench | 31.5% | 56.5% |
| lcr | 9.0% | 58.7% |
| livecodebench | 25.1% | 64.6% |
| math 500 | — | — |
| mmlu pro | 62.4% | 83.6% |
| scicode | 20.9% | 39.9% |
| tau2 | 17.3% | 54.1% |
| terminalbench hard | 2.3% | 11.4% |
Benchmark data from Artificial Analysis.