← All comparisons
Claude 4.1 Opus (Non-reasoning) vs Granite 3.3 8B (Non-reasoning)
Anthropic vs IBM — side-by-side benchmark comparison
| Claude 4.1 Opus (Non-reasoning) | Granite 3.3 8B (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 36.0 | 7.0 |
| Coding Index | — | 3.4 |
| Math Index | — | 6.7 |
| Output speed (tok/s) | 44.7 | 453.9 |
| Blended price ($/1M) | $32.81 | $0.09 |
| Time to first token (s) | 1.63s | 21.19s |
| aime | — | 4.7% |
| aime 25 | — | 6.7% |
| artificial analysis coding index | — | 3.40 |
| artificial analysis intelligence index | 36.00 | 7.00 |
| artificial analysis math index | — | 6.70 |
| gpqa | — | 33.8% |
| hle | — | 4.2% |
| ifbench | — | 22.4% |
| lcr | — | 4.3% |
| livecodebench | — | 12.7% |
| math 500 | — | 66.5% |
| mmlu pro | — | 46.8% |
| scicode | — | 10.1% |
| tau2 | — | 10.5% |
| terminalbench hard | — | 0.0% |
Benchmark data from Artificial Analysis.