← All comparisons
Olmo 3 7B Think vs DeepSeek V3.1 (Non-reasoning)
Allen Institute for AI vs DeepSeek — side-by-side benchmark comparison
| Olmo 3 7B Think | DeepSeek V3.1 (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 9.4 | 28.1 |
| Coding Index | 7.6 | 28.4 |
| Math Index | 70.7 | 49.7 |
| Output speed (tok/s) | 0.0 | 0.0 |
| Blended price ($/1M) | $0.00 | $0.83 |
| Time to first token (s) | 0.00s | 0.00s |
| aime | — | — |
| aime 25 | 70.7% | 49.7% |
| artificial analysis coding index | 7.60 | 28.40 |
| artificial analysis intelligence index | 9.40 | 28.10 |
| artificial analysis math index | 70.70 | 49.70 |
| gpqa | 51.6% | 73.5% |
| hle | 5.7% | 6.3% |
| ifbench | 41.5% | 37.8% |
| lcr | 0.0% | 45.0% |
| livecodebench | 61.7% | 57.7% |
| math 500 | — | — |
| mmlu pro | 65.5% | 83.3% |
| scicode | 21.2% | 36.7% |
| tau2 | 0.0% | 34.8% |
| terminalbench hard | 0.8% | 24.2% |
Benchmark data from Artificial Analysis.