DeepSeek v3 0324 — Benchmarks
Benchmark scores for DeepSeek v3 0324 aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.
Coding
| Benchmark | Score | Rank |
|---|---|---|
| Aider Polyglot | 55.1% | #6 / 12 |
| LiveBench Coding | 70.9 | #2 / 13 |
Math
| Benchmark | Score | Rank |
|---|---|---|
| AIME 2024/2025 | 37.8% | #8 / 22 |
| MATH Level 5 | 75.5% | #5 / 23 |
| LiveBench Math | 73.5 | #3 / 13 |
Reasoning
| Benchmark | Score | Rank |
|---|---|---|
| GPQA Diamond | 67.6% | #8 / 28 |
| LiveBench Reasoning | 65.8 | #4 / 13 |
| SimpleBench | 27.2% | #7 / 10 |
Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.