← All comparisons
Llama 3.1 Instruct 405B vs Olmo 3.1 32B Think
Meta vs Allen Institute for AI — side-by-side benchmark comparison
| Llama 3.1 Instruct 405B | Olmo 3.1 32B Think | |
|---|---|---|
| Intelligence Index | 17.4 | 13.9 |
| Coding Index | 14.5 | 9.8 |
| Math Index | 3.0 | 77.3 |
| Output speed (tok/s) | 37.5 | 0.0 |
| Blended price ($/1M) | $3.69 | $0.00 |
| Time to first token (s) | 0.63s | 0.00s |
| aime | 21.3% | — |
| aime 25 | 3.0% | 77.3% |
| artificial analysis coding index | 14.50 | 9.80 |
| artificial analysis intelligence index | 17.40 | 13.90 |
| artificial analysis math index | 3.00 | 77.30 |
| gpqa | 51.5% | 59.1% |
| hle | 4.2% | 6.0% |
| ifbench | 39.0% | 66.0% |
| lcr | 24.3% | 0.0% |
| livecodebench | 30.5% | 69.5% |
| math 500 | 70.3% | — |
| mmlu pro | 73.2% | 76.3% |
| scicode | 29.9% | 29.3% |
| tau2 | 19.0% | 0.0% |
| terminalbench hard | 6.8% | 0.0% |
Benchmark data from Artificial Analysis.