DataLearner logo

DeepSeek-V3-0324 Benchmark Details

DeepSeek-V3-0324 currently shows benchmark results led by GSM8K (3 / 26, score 96.30), GPQA (3 / 15, score 68.40), DROP (3 / 9, score 89.70).

Benchmark Results

DeepSeek-V3-0324

Benchmark Results

Thinking

General Knowledge

6 evaluations
Benchmark / mode
Score
Rank/total
MMLU
86.50
28 / 66
MMLU Pro
81.20
55 / 132
GPQA Diamond
68.40
125 / 187
GPQA
68.40
3 / 15
ARC-AGI
9
62 / 68
HLE
5.20
165 / 172

Math and Reasoning

7 evaluations
Benchmark / mode
Score
Rank/total
GSM8K
96.30
3 / 26
MATH-500
94
28 / 44
AIME 2024
59.40
43 / 62
AIME2025
47.70
89 / 107
IMO-ProofBench
4.30
15 / 16
IMO 2024
1.70
9 / 10
IMO 2025
1.70
9 / 9

Reading Comprehension

1 evaluations
Benchmark / mode
Score
Rank/total
DROP
89.70
3 / 9

Common Sense

1 evaluations
Benchmark / mode
Score
Rank/total
SimpleQA
27.20
28 / 47

Coding and Software Engineer

2 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
49.20
95 / 123
SWE-bench Verified
38.80
103 / 112

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
81.60
15 / 23

AI Agent - Tool Usage

1 evaluations
Benchmark / mode
Score
Rank/total
Terminal-Bench
13.30
34 / 35

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
Simple Bench
Standard Mode
27.20
51 / 63

Agent Level Benchmark

2 evaluations
Benchmark / mode
Score
Rank/total
Aider-Polyglot
Standard Mode
55.10
27 / 59
τ²-Bench
38.80
38 / 43
Compare with other models