DataLearner logo

DeepSeek-R1 Benchmark Details

DeepSeek-R1 currently shows benchmark results led by MMLU (8 / 66, score 90.80), MMLU Pro (38 / 132, score 84), MATH-500 (13 / 44, score 97.30).

Benchmark Results

DeepSeek-R1

Benchmark Results

Thinking

General Knowledge

4 evaluations
Benchmark / mode
Score
Rank/total
MMLU
90.80
8 / 66
MMLU Pro
84
38 / 132
GPQA Diamond
71.50
109 / 187
ARC-AGI
15.80
58 / 68

Common Sense

1 evaluations
Benchmark / mode
Score
Rank/total
SimpleQA
30.10
24 / 47

Coding and Software Engineer

2 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
65.90
60 / 123
SWE-bench Verified
49.20
96 / 112

Math and Reasoning

3 evaluations
Benchmark / mode
Score
Rank/total
MATH-500
97.30
13 / 44
AIME 2024
79.80
28 / 62
AIME2025
70
74 / 107

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
84.60
11 / 23

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
Simple Bench
Standard Mode
30.90
48 / 63

Agent Level Benchmark

1 evaluations
Benchmark / mode
Score
Rank/total
Aider-Polyglot
Thinking Enabled
56.90
25 / 59
Compare with other models