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

BenchmarkScoreRank
Aider Polyglot55.1%#6 / 12
LiveBench Coding70.9#2 / 13

Math

BenchmarkScoreRank
AIME 2024/202537.8%#8 / 22
MATH Level 575.5%#5 / 23
LiveBench Math73.5#3 / 13

Reasoning

BenchmarkScoreRank
GPQA Diamond67.6%#8 / 28
LiveBench Reasoning65.8#4 / 13
SimpleBench27.2%#7 / 10

Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.