benchgap
DeepSeek · model

DeepSeek-R1-0528-Qwen3-8B benchmark scores

As of 2026-11-19, DeepSeek-R1-0528-Qwen3-8B (DeepSeek) has measured scores on 5 benchmarks and estimated scores on 24 more.

BenchmarkScoreSource
AA-GPQA Diamond59.0%estimated ± 6.2 pp, medium confidence
AA-HLE5.0%estimated ± 5.7 pp, medium confidence
AIME2676.9%estimated ± 4.3 pp, medium confidence
Artificial Analysis Intelligence Index6.7%estimated ± 7.4 pp, medium confidence
CritPt0.3%estimated ± 4.0 pp, high confidence
GPQA Diamond61.4%estimated ± 1.2 pp, high confidence
HLE5.2%estimated ± 6.5 pp, low confidence
AA-Omniscience Accuracy19.1%estimated ± 8.0 pp, medium confidence
SWE-bench Verified58.0%estimated ± 9.3 pp, low confidence
MMLU-Pro76.8%estimated ± 4.4 pp, high confidence
AIME 202486.0%measured
BBH82.7%estimated ± 4.5 pp, low confidence
GPQA61.1%measured
MMLU87.7%estimated ± 4.2 pp, high confidence
Vals GPQA Diamond59.1%estimated ± 4.5 pp, high confidence
Vals MMLU-Pro74.6%estimated ± 2.3 pp, high confidence
MATH-50097.5%estimated ± 1.2 pp, high confidence
MMMLU78.9%estimated ± 1.7 pp, high confidence
AIME 202576.3%measured
AA MMLU-Pro84.4%estimated ± 1.2 pp, low confidence
HLE w/o tools11.6%estimated ± 5.1 pp, low confidence
C-Eval85.9%estimated ± 1.2 pp, high confidence
HMMT Feb 202561.5%measured
HMMT Feb 202642.9%estimated ± 11.0 pp, low confidence
HMMT Nov 202564.2%estimated ± 1.5 pp, medium confidence
MMLU-Redux87.9%estimated ± 1.7 pp, high confidence
SuperGPQA42.2%estimated ± 8.5 pp, low confidence
SimpleQA14.6%estimated ± 13.4 pp, low confidence
LiveCodeBench60.5%measured