← All comparisons
DeepSeek V4 Pro (Non-reasoning) vs MiMo-V2-Omni-0327
DeepSeek vs Xiaomi — side-by-side benchmark comparison
| DeepSeek V4 Pro (Non-reasoning) | MiMo-V2-Omni-0327 | |
|---|---|---|
| Intelligence Index | 39.3 | 44.9 |
| Coding Index | 38.4 | 36.9 |
| Math Index | — | — |
| Output speed (tok/s) | 53.3 | 91.7 |
| Blended price ($/1M) | $0.54 | $0.80 |
| Time to first token (s) | 1.18s | 2.18s |
| aime | — | — |
| aime 25 | — | — |
| artificial analysis coding index | 38.40 | 36.90 |
| artificial analysis intelligence index | 39.30 | 44.90 |
| artificial analysis math index | — | — |
| gpqa | 71.7% | 85.5% |
| hle | 7.7% | 20.4% |
| ifbench | 45.8% | 67.3% |
| lcr | 45.0% | 63.7% |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | — |
| scicode | 42.4% | 39.5% |
| tau2 | 91.2% | 88.0% |
| terminalbench hard | 36.4% | 35.6% |
Benchmark data from Artificial Analysis.