Anthropic

Claude Sonnet 4.6

by Anthropictext+image+file->text9 endpoints
Neura Intelligence Index
75.7/ 100
Rank
#55
Confidence
high
Benchmarks
9(23% coverage)

Overview

Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work. It excels at iterative development, complex codebase navigation, end-to-end project management with...

Capabilities

Text generation
Image understanding
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
File input
Top-P sampling
Stop sequences

Modalities

Input
TextImageFile
Output
Text

Technical Specifications

Context Window
1M tokens
Max Output
128K tokens
Tokenizer
Claude
Content Moderation
Enabled
Uptime (24h)
10000.0%
Country
🇺🇸US
Release Date
February 17, 20267mo

Supported Parameters

include_reasoningmax_completion_tokensmax_tokensreasoningreasoning_effortresponse_formatstopstructured_outputstemperaturetool_choicetoolstop_ktop_pverbosity

Benchmark Performance

general

BenchmarkScoreSource
MMMLU
89.3
verified

coding

BenchmarkScoreSource
SWE-bench Verified
79.6
verified

reasoning

BenchmarkScoreSource
GPQA Diamond
89.9
verified
ARC-AGI v2
58.3
verified
Humanity's Last Exam
49
verified

multimodal

BenchmarkScoreSource
MMMU-Pro
75.6
verified

agent

BenchmarkScoreSource
BrowseComp
74.7
verified
OSWorld
72.5
verified
MCP Atlas
61.3
verified

Performance Over Time

Pricing

Per 1M tokens
Input$3.00
Output$15.00
Cache read$0.30
Cache write$3.75
Added
February 17, 2026
Last synced 9/21/2026