benchgap
DeepSeek · model

DeepSeek V3 benchmark scores

As of 2026-11-21, DeepSeek V3 (DeepSeek) has measured scores on 21 benchmarks and estimated scores on 76 more.

BenchmarkScoreSource
AA-GPQA Diamond55.7%measured
AA-HLE2.9%measured
AA-SciCode35.8%measured
Artificial Analysis Intelligence Index8.5%measured
CritPt0.0%measured
GDPval-AA0.0%measured
GPQA Diamond59.4%estimated ± 1.2 pp, high confidence
HLE4.6%estimated ± 6.5 pp, low confidence
IFBench53.8%estimated ± 6.6 pp, low confidence
AA-LCR29.3%measured
LiveCodeBench v673.9%estimated ± 4.7 pp, high confidence
AA-Omniscience Accuracy25.5%measured
SciCode36.1%estimated ± 3.1 pp, high confidence
BrowseComp7.7%estimated ± 9.0 pp, low confidence
HLE w/ tools33.8%estimated ± 5.3 pp, low confidence
IFEval86.1%measured
LongBench v248.7%measured
τ²-bench results22.8%measured
AA Coding Index23.0%measured
APEX-Agents0.0%estimated ± 9.4 pp, low confidence
APEX-Agents-AA6.0%estimated ± 9.5 pp, medium confidence
BioMysteryBench (human-difficult)0.0%estimated ± 6.8 pp, low confidence
DeepSearchQA17.0%estimated ± 4.5 pp, medium confidence
SWE-bench Verified42.0%measured
BFCL v453.9%estimated ± 2.7 pp, high confidence
CyberGym46.2%estimated ± 5.3 pp, low confidence
JobBench0.0%estimated ± 8.7 pp, low confidence
SWE-bench Pro16.4%estimated ± 4.5 pp, medium confidence
WideResearch0.0%estimated ± 7.8 pp, low confidence
AA Agentic Index0.8%measured
MMLU-Pro75.9%measured
AA-IFBench34.8%measured
AIME 202439.2%measured
BBH93.4%estimated ± 2.8 pp, medium confidence
GPQA59.1%measured
MMLU90.5%estimated ± 0.9 pp, high confidence
Vals GPQA Diamond54.8%estimated ± 3.2 pp, high confidence
Vals LiveCodeBench73.2%estimated ± 5.7 pp, low confidence
Vals MMLU-Pro73.5%estimated ± 2.3 pp, high confidence
MATH-50090.2%measured
MMMLU83.3%estimated ± 0.7 pp, medium confidence
AIME 202525.6%estimated ± 3.5 pp, high confidence
AA MMLU-Pro84.9%estimated ± 0.8 pp, high confidence
Gert Labs33.0%estimated ± 8.3 pp, low confidence
AA Harvey LAB v1.079.4%estimated ± 1.2 pp, low confidence
AA Tau3 Banking16.5%estimated ± 3.7 pp, medium confidence
ARC-AGI-174.7%estimated ± 7.8 pp, low confidence
ARC-AGI-230.5%estimated ± 14.4 pp, low confidence
CursorBench 3.231.0%estimated ± 5.3 pp, low confidence
FrontierSWE v22.9%estimated ± 14.2 pp, low confidence
MazeBench0.4%estimated ± 5.5 pp, low confidence
OSWorld-Verified55.4%estimated ± 8.4 pp, low confidence
Vals SWE-bench47.0%estimated ± 5.5 pp, medium confidence
CWE-Bench-AA0.0%estimated ± 10.1 pp, low confidence
AA Cyber Index0.6%estimated ± 9.2 pp, low confidence
DeepsecBench-AA0.0%estimated ± 8.9 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)48.4%estimated ± 4.3 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)38.3%estimated ± 1.1 pp, low confidence
AA ITBench33.1%estimated ± 4.0 pp, medium confidence
Bug Hunt Bench12.1%estimated ± 10.5 pp, low confidence
CWE-bench v10.0%estimated ± 6.4 pp, low confidence
DeepSWE20.3%estimated ± 7.5 pp, low confidence
DRACO0.0%estimated ± 2.0 pp, low confidence
HLE w/o tools22.6%estimated ± 4.9 pp, high confidence
PostTrainBench v1.124.9%estimated ± 7.4 pp, low confidence
Toolathlon-Verified3.2%estimated ± 2.3 pp, low confidence
VulcanBench Frontier v464.8%estimated ± 6.8 pp, low confidence
GDP.pdf17.0%estimated ± 4.6 pp, medium confidence
ARC-AGI-31.3%estimated ± 9.8 pp, low confidence
ARC-AGI-3 Provider Adapter2.7%estimated ± 7.3 pp, low confidence
Vibe Code Bench0.0%estimated ± 8.4 pp, low confidence
Hermes Bench0.0%estimated ± 2.9 pp, medium confidence
Hermes Index0.0%estimated ± 8.7 pp, low confidence
C-Eval88.1%estimated ± 1.0 pp, low confidence
MCP Atlas64.6%estimated ± 7.8 pp, low confidence
MMLU-Redux87.9%estimated ± 0.6 pp, high confidence
NL2Repo14.0%estimated ± 4.7 pp, medium confidence
QwenClawBench55.9%estimated ± 6.0 pp, low confidence
SuperGPQA41.1%estimated ± 8.5 pp, low confidence
Toolathlon27.0%estimated ± 6.3 pp, low confidence
AA LiveCodeBench84.0%estimated ± 4.7 pp, low confidence
HealthBench Hard0.0%estimated ± 8.6 pp, low confidence
MMLU-Pro (Arcee)77.1%estimated ± 4.0 pp, low confidence
React Native Evals62.5%estimated ± 2.7 pp, low confidence
ResearchClawBench1.6%estimated ± 1.9 pp, medium confidence
BioMysteryBench (human-solvable)81.7%estimated ± 0.6 pp, low confidence
HealthBench (length-adjusted)54.1%estimated ± 3.7 pp, medium confidence
HealthBench Professional56.7%estimated ± 4.2 pp, low confidence
HLE-Verified20.3%estimated ± 7.1 pp, low confidence
ScarfBench24.4%estimated ± 9.1 pp, low confidence
STEMBench2.4%estimated ± 0.2 pp, low confidence
SimpleQA24.9%measured
LiveCodeBench37.6%measured
OpenHarmony Bench v1.051.0%estimated ± 2.4 pp, medium confidence
MRCRv222.8%estimated ± 11.3 pp, low confidence
MMLongBench-Doc56.0%estimated ± 2.6 pp, low confidence
BioBasicsBench34.1%estimated ± 3.8 pp, low confidence