Claude Sonnet 5 Benchmark Details

Claude Sonnet 5 currently shows benchmark results led by HLE (9 / 181, score 57.40), SWE-bench Verified (7 / 114, score 85.20), BrowseComp (7 / 54, score 84.70). 1 source link is attached for reference.

Benchmark Results

Claude Sonnet 5

Benchmark Results

General Knowledge

2 evaluations

Benchmark / mode

Score

Rank/total

HLE

Thinking Level · Extra High

43.20

52 / 181

HLE

Thinking Level · Extra HighTools

57.40

9 / 181

Other

2 evaluations

Benchmark / mode

Score

Rank/total

GPQA Diamond

Thinking Level · Max

80.30

117 / 224

GPQA Diamond

Thinking Level · Extra High

90.53

33 / 224

Coding and Software Engineer

4 evaluations

Benchmark / mode

Score

Rank/total

Text Arena (Coding)

Thinking Level · High

1544.16

10 / 35

SWE-bench Verified

Thinking Level · Extra HighTools

85.20

7 / 114

WeirdML v2

Thinking Level · HighTools

68.78

20 / 52

DeepSWE

DeepTools

17 / 28

Writing and Creative Capabilities

1 evaluations

Benchmark / mode

Score

Rank/total

Creative Writing

Standard Mode

1787.60

18 / 99

Common Sense Reasoning

1 evaluations

Benchmark / mode

Score

Rank/total

SimpleBench

Standard Mode

57.90

23 / 67

AI Agent - Information Search

1 evaluations

Benchmark / mode

Score

Rank/total

BrowseComp

Thinking EnabledToolsInternet

84.70

7 / 54

AI Agent - Tool Usage

3 evaluations

Benchmark / mode

Score

Rank/total

OSWorld-Verified

Thinking Level · Extra HighTools

81.20

6 / 26

Terminal-Bench 2.1

Thinking Level · HighTools

74.60

27 / 45

Terminal-Bench 2.1

Thinking Level · Extra HighTools

80.40

19 / 45

Math and Reasoning

2 evaluations

Benchmark / mode

Score

Rank/total

FrontierMath v2

Thinking Level · Max

65.61

17 / 34

FrontierMath Tier 4 v2

Thinking Level · Max

29.27

16 / 34

Compare with other models

Sources

anthropic.comanthropic.com