DeepSeek-R1-0528 Benchmark Details

DeepSeek-R1-0528 currently shows benchmark results led by MMLU-Pro (28 / 175, score 85), MATH-500 (8 / 46, score 98), AIME 2024 (13 / 61, score 91.40).

Benchmark Results

DeepSeek-R1-0528

Benchmark Results

Knowledge Exams

2 evaluations

Benchmark / mode

Score

Rank/total

MMLU-Pro

Thinking Mode

28 / 175

HLE

Thinking Mode

17.70

169 / 235

Abstract Generalization

2 evaluations

Benchmark / mode

Score

Rank/total

ARC-AGI-1

Thinking Mode

21.20

157 / 176

ARC-AGI-2

Thinking Mode

1.30

151 / 164

Scientific Reasoning

2 evaluations

Benchmark / mode

Score

Rank/total

GPQA Diamond

Thinking Mode

122 / 253

CritPt

Thinking Mode

1.40

139 / 204

Honesty & Factuality

1 evaluations

Benchmark / mode

Score

Rank/total

SimpleQA

Thinking Mode

27.80

27 / 46

Repository Engineering

1 evaluations

Benchmark / mode

Score

Rank/total

SWE-bench Verified

Thinking Mode

57.60

86 / 116

Mathematics

5 evaluations

Benchmark / mode

Score

Rank/total

MATH-500

Thinking Mode

8 / 46

AIME 2024

Thinking Mode

91.40

13 / 61

AIME2025

Thinking Mode

87.50

49 / 110

IMO-ProofBench

Thinking Mode

7 / 16

IMO-ProofBench Advanced

Thinking Mode

3.80

22 / 24

Algorithmic Coding

1 evaluations

Benchmark / mode

Score

Rank/total

LiveCodeBench

Thinking Mode

73.30

49 / 125

Writing

1 evaluations

Benchmark / mode

Score

Rank/total

Creative Writing

Standard Mode

1421.10

63 / 110

Agentic Development

3 evaluations

Benchmark / mode

Score

Rank/total

Aider-Polyglot

Thinking Mode

71.40

15 / 59

Terminal Bench Hard

Thinking ModeTools

15.90

165 / 244

Terminal-Bench

Thinking Mode

5.70

35 / 35

Commonsense

1 evaluations

Benchmark / mode

Score

Rank/total

SimpleBench

Thinking Mode

40.80

68 / 96

Capability Indices

1 evaluations

Benchmark / mode

Score

Rank/total

ECI

unknown

141.30

91 / 167

Compare with other models