DeepSeek v3 — Benchmarks
Benchmark scores for DeepSeek v3 aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.
Overall rank: #33 of 73 open modelscomposite 50/100 across 11 benchmarks in 4 categories · methodology
Coding
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| Aider Polyglot | 48.4% | #9 / 18 | #40 / 69 |
| SWE-bench Lite | 36.7% | #2 / 3 | #34 / 80 |
Knowledge
Math
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| AIME 2024/2025 | 15.8% | #18 / 34 | #113 / 155 |
| MATH Level 5 | 64.8% | #10 / 32 | #50 / 108 |
| FrontierMath | 1.7% | #10 / 12 | #85 / 101 |
Reasoning
| Benchmark | Score | Open rank | All models |
|---|---|---|---|
| GPQA Diamond | 56.5% | #16 / 46 | #109 / 182 |
| BIG-Bench Hard | 87.5% | #1 / 37 | #2 / 50 |
| SimpleBench | 18.9% | #19 / 19 | #85 / 90 |
Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.