DataLearner logo

Claude Sonnet 4 Benchmark Details

Claude Sonnet 4 currently shows benchmark results led by SWE-bench Verified (14 / 113, score 80.20), Terminal-Bench (10 / 35, score 41.30), MMLU Pro (38 / 132, score 84). 1 source link is attached for reference.

Benchmark Results

Claude Sonnet 4

Benchmark Results

Thinking
Tool usage

General Knowledge

12 evaluations
Benchmark / mode
Score
Rank/total
MMLU Pro
84
38 / 132
GPQA Diamond
83.80
62 / 188
GPQA Diamond
75.40
98 / 188
GPQA Diamond
68
129 / 188
LiveBench
Standard Mode
50.98
89 / 115
LiveBench
64K
61.27
65 / 115
ARC-AGI
40
49 / 68
ARC-AGI
23.80
56 / 68
HLE
9.60
150 / 173
HLE
5.52
164 / 173
ARC-AGI-2
5.90
46 / 62
ARC-AGI-2
1.30
55 / 62

Coding and Software Engineer

6 evaluations
Benchmark / mode
Score
Rank/total
CodeClash
Standard ModeTools
1223
4 / 8
SWE-bench Verified
80.20
14 / 113
SWE-bench Verified
72.70
52 / 113
LiveCodeBench
66
59 / 123
LiveCodeBench
48.50
96 / 123
SWE-Bench Pro - Public
42.70
49 / 55

Math and Reasoning

12 evaluations
Benchmark / mode
Score
Rank/total
AIME2025
85
51 / 107
AIME2025
70.50
72 / 107
AIME2025
38
96 / 107
AIME 2024
43.40
50 / 62
IMO-ProofBench
27.10
8 / 16
IMO 2024
9.70
5 / 10
IMO 2024
5.20
8 / 10
IMO-ProofBench Advanced
4.80
6 / 8
FrontierMath
4.10
41 / 60
IMO 2025
4
5 / 9
IMO 2025
3.30
6 / 9
FrontierMath - Tier 4
Standard Mode
0
72 / 80

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
83.05
14 / 23

AI Agent - Tool Usage

4 evaluations
Benchmark / mode
Score
Rank/total
OSWorld-Verified
42.20
23 / 25
Terminal-Bench
41.30
10 / 35
Terminal-Bench
35.50
18 / 35
Terminal-Bench
26
26 / 35

Multimodal Understanding

1 evaluations
Benchmark / mode
Score
Rank/total
MMMU
76.50
17 / 29

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
Simple Bench
Thinking Enabled
45.50
34 / 63

Agent Level Benchmark

4 evaluations
Benchmark / mode
Score
Rank/total
τ²-Bench - Telecom
65
29 / 35
Aider-Polyglot
Standard Mode
56.40
26 / 59
Aider-Polyglot
32K
61.30
20 / 59
τ²-Bench
52
34 / 43

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
55
24 / 31

Productivity Knowledge

1 evaluations
Benchmark / mode
Score
Rank/total
GDPval-AA
33
19 / 21

Long Context

1 evaluations
Benchmark / mode
Score
Rank/total
AA-LCR
65
13 / 16

Claw-style Agent Evaluation

2 evaluations
Benchmark / mode
Score
Rank/total
Pinch Bench
Thinking EnabledTools
80.50
22 / 37
Claw Bench
Thinking EnabledTools
77.80
23 / 29
Compare with other models

Sources

artificialanalysis.aiartificialanalysis.ai