← All comparisons
Granite 4.1 8B vs Jamba 1.7 Large
IBM vs AI21 Labs — side-by-side benchmark comparison
| Granite 4.1 8B | Jamba 1.7 Large | |
|---|---|---|
| Intelligence Index | 12.4 | 10.9 |
| Coding Index | 7.3 | 7.8 |
| Math Index | — | 2.3 |
| Output speed (tok/s) | 111.5 | 56.9 |
| Blended price ($/1M) | $0.06 | $3.50 |
| Time to first token (s) | 0.41s | 0.90s |
| aime | — | 5.7% |
| aime 25 | — | 2.3% |
| artificial analysis coding index | 7.30 | 7.80 |
| artificial analysis intelligence index | 12.40 | 10.90 |
| artificial analysis math index | — | 2.30 |
| gpqa | 43.3% | 39.0% |
| hle | 3.8% | 3.8% |
| ifbench | 38.6% | 35.2% |
| lcr | 12.0% | 17.3% |
| livecodebench | — | 18.1% |
| math 500 | — | 60.0% |
| mmlu pro | — | 57.7% |
| scicode | 21.8% | 18.8% |
| tau2 | 27.8% | 13.5% |
| terminalbench hard | 0.0% | 2.3% |
Benchmark data from Artificial Analysis.