← All comparisons
Olmo 3 7B Think vs Llama 3.1 Instruct 70B
Allen Institute for AI vs Meta — side-by-side benchmark comparison
| Olmo 3 7B Think | Llama 3.1 Instruct 70B | |
|---|---|---|
| Intelligence Index | 9.4 | 12.5 |
| Coding Index | 7.6 | 10.9 |
| Math Index | 70.7 | 4.0 |
| Output speed (tok/s) | 0.0 | 35.3 |
| Blended price ($/1M) | $0.00 | $0.56 |
| Time to first token (s) | 0.00s | 0.54s |
| aime | — | 17.3% |
| aime 25 | 70.7% | 4.0% |
| artificial analysis coding index | 7.60 | 10.90 |
| artificial analysis intelligence index | 9.40 | 12.50 |
| artificial analysis math index | 70.70 | 4.00 |
| gpqa | 51.6% | 40.9% |
| hle | 5.7% | 4.6% |
| ifbench | 41.5% | 34.4% |
| lcr | 0.0% | 6.3% |
| livecodebench | 61.7% | 23.2% |
| math 500 | — | 64.9% |
| mmlu pro | 65.5% | 67.6% |
| scicode | 21.2% | 26.7% |
| tau2 | 0.0% | 15.2% |
| terminalbench hard | 0.8% | 3.0% |
Benchmark data from Artificial Analysis.