← All comparisons
DeepSeek V3.1 Terminus (Reasoning) vs Seed-OSS-36B-Instruct
DeepSeek vs ByteDance Seed — side-by-side benchmark comparison
| DeepSeek V3.1 Terminus (Reasoning) | Seed-OSS-36B-Instruct | |
|---|---|---|
| Intelligence Index | 33.9 | 25.2 |
| Coding Index | 33.7 | 16.7 |
| Math Index | 89.7 | 84.7 |
| Output speed (tok/s) | 0.0 | 44.4 |
| Blended price ($/1M) | $1.91 | $0.30 |
| Time to first token (s) | 0.00s | 1.88s |
| aime | — | — |
| aime 25 | 89.7% | 84.7% |
| artificial analysis coding index | 33.70 | 16.70 |
| artificial analysis intelligence index | 33.90 | 25.20 |
| artificial analysis math index | 89.70 | 84.70 |
| gpqa | 79.2% | 72.6% |
| hle | 15.2% | 9.1% |
| ifbench | 57.0% | 41.9% |
| lcr | 65.0% | 57.7% |
| livecodebench | 79.8% | 76.5% |
| math 500 | — | — |
| mmlu pro | 85.1% | 81.5% |
| scicode | 40.6% | 36.5% |
| tau2 | 37.1% | 49.4% |
| terminalbench hard | 30.3% | 6.8% |
Benchmark data from Artificial Analysis.