Claude 4.6 Sonnet

Closed Source

Closed Weights

Parameters

Context Length

200K

Modality

Multimodal

Architecture

Dense

License

Proprietary

Release Date

1 Feb 2026

Knowledge Cutoff

Evaluation Benchmarks

Rank

#24

Benchmark	Score	Rank
StackUnseen ProLLM Stack Unseen	0.891	🥈 2
Coding medium LiveBench Coding	0.79	20
Data Analysis auto LiveBench Data Analysis	0.78	22
Data Analysis medium LiveBench Data Analysis	0.78	22
Web Development high WebDev Arena	1546	26
Web Development Standard WebDev Arena	1522	31
Reasoning medium LiveBench Reasoning	0.85	34
General Text Text Arena	1472	38
General medium LiveBench Global	0.73	39
LiveBench Average medium LiveBench Average	0.73	45
Mathematics medium LiveBench Mathematics	0.87	47
Agentic Coding medium LiveBench Agentic	0.43	52

Rankings

Overall Rank

#24

Coding Rank

#12

About Claude 4.6 Sonnet

Claude 4.6 Sonnet delivers the optimal balance of intelligence, speed, and cost for enterprise applications. Features enhanced reasoning, coding capabilities, and multimodal understanding with vision support. Excels at complex tasks including data analysis, content generation, code review, and conversational AI. Offers improved accuracy and reliability over previous generations with constitutional AI safety measures.

Technical Specifications

Detailed architecture specifications are limited for proprietary models.

Resources

Official Documentation

About Claude 4.6

Anthropic's Claude 4.6 series introduces breakthrough capabilities in extended reasoning, creative collaboration, and safety. Features variants including Opus Thinking with advanced chain-of-thought processing and Sonnet for balanced performance. These models excel at complex reasoning tasks, coding, creative writing, and nuanced analysis with enhanced constitutional AI safeguards.

Other Claude 4.6 Models

No related models available