DeepSeek V4.1 Flash — Benchmarks

Benchmark scores for DeepSeek V4.1 Flash aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.

Overall rank: #4 of 78 open modelscomposite 65.8/100 across 8 benchmarks in 3 categories · methodology

Coding

BenchmarkScoreOpen rankAll models
LiveBench Coding80.0#2 / 19#19 / 63
SciCode51.8%#4 / 62#53 / 161
LMArena WebDev1616.1#5 / 48#17 / 123
Surface Evolver Bench46.3%#6 / 13#19 / 27

Math

BenchmarkScoreOpen rankAll models
ProofBench54.0%#4 / 24#17 / 68
LiveBench Math93.3#2 / 19#18 / 63

Reasoning

BenchmarkScoreOpen rankAll models
DTBench89.9%#8 / 67#71 / 210
LMCA47.0%#3 / 53#63 / 172
CritPt14.3%#7 / 62#58 / 172
LiveBench Reasoning86.7#3 / 19#27 / 63

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