Claude Sonnet 4.6: Frontier Coding at Sonnet Pricing

What's New in Claude Sonnet 4.6

Claude Sonnet 4.6 is Anthropic's most capable Sonnet model, with major upgrades in coding, computer use, long-context reasoning, and agent planning. It is now the default model on claude.ai for Free and Pro users.

Developers with early access prefer Sonnet 4.6 over its predecessor by a wide margin, and often even over Claude Opus 4.5, Anthropic's smartest model from November 2025.

Claude Sonnet 4.6 Coding Performance

Performance that previously required an Opus-class model is now available at Sonnet pricing ($3/$15 per million tokens). Key improvements:

Better code generation: More consistent, accurate code output across languages
Improved instruction following: Follows complex multi-step coding instructions more precisely
Stronger debugging: Better at catching its own mistakes and suggesting fixes
Real-world task performance: State-of-the-art on economically valuable office tasks (GDPval-AA)

For teams using AI-powered security scanning, Claude Code Security works with both Sonnet 4.6 and Opus 4.6 to detect vulnerabilities in codebases.

Computer Use Benchmarks: OSWorld Results

Anthropic pioneered general-purpose computer use in October 2024. On OSWorld, the standard benchmark where AI completes tasks across real software like Chrome, VS Code, and LibreOffice, Sonnet models have shown steady improvement over 16 months:

Model	OSWorld Score
Sonnet 3.5 (Oct 2024)	Baseline
Sonnet 3.6	+15%
Sonnet 4.5	+40%
Sonnet 4.6	+65%

Early users report human-level capability on tasks like navigating complex spreadsheets, filling out multi-step web forms, and working across multiple browser tabs.

Prompt Injection Resistance

Computer use poses security risks from prompt injection attacks on websites. Sonnet 4.6 shows a major improvement in injection resistance compared to Sonnet 4.5, performing similarly to the more expensive Opus 4.6.

1M Token Context Window

Sonnet 4.6 features a 1M token context window in beta, enough to process entire codebases, long documents, or extensive conversation histories in a single request.

What Claude Sonnet 4.6 Means for Developers

For developers, Sonnet 4.6 represents a significant cost-efficiency improvement. Tasks that previously needed Opus-class models (at $5/$25 per million tokens) now perform comparably at Sonnet pricing. This makes AI-powered development more accessible for:

Agentic coding workflows: Longer, more reliable automated coding sessions
Code review and debugging: Catch issues before they reach production
Computer use automation: Automate legacy software interactions
Large codebase analysis: Use the 1M context window to understand entire projects

Frequently Asked Questions

What is Claude Sonnet 4.6?

Claude Sonnet 4.6 is Anthropic's most capable Sonnet-tier model, released February 2026. It delivers coding and reasoning performance that previously required Opus-class models, but at Sonnet pricing ($3/$15 per million tokens). It is now the default model on claude.ai for Free and Pro users and includes a 1M token context window in beta.

How much does Claude Sonnet 4.6 cost?

Claude Sonnet 4.6 costs $3 per million input tokens and $15 per million output tokens, the same as Sonnet 4.5. This is 40% cheaper than Opus pricing ($5/$25). It is available on claude.ai, the Anthropic API with model ID claude-sonnet-4-6, Amazon Bedrock, and Google Cloud Vertex AI.

Is Claude Sonnet 4.6 better than Opus 4.5 for coding?

Yes. Developers with early access frequently preferred Sonnet 4.6 over Claude Opus 4.5 for coding tasks, despite Sonnet being a cheaper tier. Sonnet 4.6 shows particular strength in code generation, instruction following, and debugging. For the most demanding tasks, Claude Opus 4.6 still leads on benchmarks like Terminal-Bench 2.0.

What is Claude Sonnet 4.6 computer use?

Computer use allows Claude to interact with software like a human, clicking buttons, filling forms, and navigating UIs. On OSWorld, the standard benchmark for computer use, Sonnet 4.6 scores 65% higher than the original Sonnet 3.5 baseline from October 2024. It also has significantly improved prompt injection resistance, performing similarly to the more expensive Opus 4.6.