Claude Sonnet 4.6
anthropic/claude-sonnet-4-6
Claude workhorse for coding agents, careful analysis, and production cost control
Capabilities
ReasoningYes
Tool callingYes
Structured outputNo
TemperatureSupported
AttachmentYes
Open weightsNo
Limits
Context1,000,000 tokens
Max inputN/A
Max output64,000 tokens
Details
Familyclaude-sonnet
Knowledge cutoff2025-08-31
Release date2026-02-17
Last updated2026-03-13
Modalities
Inputtext, image, pdf
Outputtext
Benchmarks
SWE-Atlas Codebase QnA
31.2
score
SWE-Atlas Refactoring
32.21
score
SWE-Atlas Test Writing
31.76
score
Artificial Analysis Coding Agent Index
49.4
average pass@1
SWE-Atlas Codebase QnA
70.3
pass@1
SWE-Bench Pro
14.9
pass@1
Terminal-Bench
63.1
pass@1
Terminal-Bench
67
success rate
Humanity's Last Exam
34.6
accuracy
Humanity's Last Exam
46.8
accuracy
OSWorld-Verified
78.5
success rate
Community Rating
—
No ratings yet
Sign in to rate
Pricing
Cost per 1M tokens
InputN/A
OutputN/A