Claude Sonnet 4.6
Anthropic
Parameters
—
Architecture
—
Released
—
License
—
About
No detailed description available for this model.
Benchmark Scores
| Benchmark | Score | Date | Standing |
|---|---|---|---|
|
YC-Bench
general_agent
|
4.55%
|
11.06.2026 |
#7/8
better than 14%
|
|
Claw-Eval Avg
coding_agent
|
82.31%
|
11.06.2026 |
#13/40
better than 69%
|
|
OSWorld-Verified
general_agent
|
75.00%
|
11.06.2026 |
#13/25
better than 50%
|
|
OmniDocBench 1.5
document_understanding
|
94.86%
|
11.06.2026 |
#12/26
better than 56%
|
|
AgentWorldBench MCP
general_agent
|
99.34%
|
24.06.2026 |
#2/14
better than 92%
|
|
AgentWorldBench Search
general_agent
|
43.14%
|
24.06.2026 |
#8/14
better than 46%
|
|
AgentWorldBench Terminal
general_agent
|
87.47%
|
24.06.2026 |
#4/14
better than 77%
|
|
AgentWorldBench SWE
general_agent
|
87.21%
|
24.06.2026 |
#5/14
better than 69%
|
|
AgentWorldBench Android
general_agent
|
60.28%
|
24.06.2026 |
#9/14
better than 38%
|
|
AgentWorldBench Web
general_agent
|
48.68%
|
24.06.2026 |
#7/14
better than 46%
|
|
AgentWorldBench OS
general_agent
|
49.53%
|
24.06.2026 |
#8/14
better than 46%
|
|
AgentWorldBench Overall
general_agent
|
78.79%
|
24.06.2026 |
#6/14
better than 62%
|
|
SWE-bench Verified
coding_agent
|
90.83%
|
11.06.2026 |
#11/60
better than 83%
|
|
SWE Atlas - QnA
coding_agent
|
40.15%
|
11.06.2026 |
#16/24
better than 35%
|
|
SWE Atlas - TW
coding_agent
|
42.61%
|
11.06.2026 |
#11/20
better than 47%
|
|
VIBE-V2
coding_agent
|
53.05%
|
11.06.2026 |
#7/9
better than 25%
|
|
SVG-Bench
coding_agent
|
100.00%
|
11.06.2026 |
#1/9
better than 100%
|
|
DRACO
general_agent
|
45.30%
|
11.06.2026 |
#9/12
better than 27%
|
|
GDPVal
general_agent
|
81.87%
|
11.06.2026 |
#4/20
better than 84%
|
|
MMMU-Pro
vision_language
|
83.36%
|
11.06.2026 |
#19/35
better than 47%
|
|
Apex-Agents
general_agent
|
32.09%
|
11.06.2026 |
#20/24
better than 17%
|
|
BrowseComp
general_agent
|
80.45%
|
11.06.2026 |
#15/45
better than 68%
|
|
MCP-Atlas
general_agent
|
63.58%
|
11.06.2026 |
#40/54
better than 26%
|
What is a good YC-Bench score? What is a good Claw-Eval Avg score? What is a good OSWorld-Verified score?
No comments yet
Be the first to share your thoughts!