Claude Sonnet 4
by Anthropic
Balanced hybrid reasoning model released May 2025, excelling in coding with 72.7% SWE-bench performance. Features dual modes for instant responses or extended thinking, making it ideal for practical AI applications requiring both speed and intelligence.
- Hybrid reasoning model with instant responses and extended thinking modes for optimal task handling
- State-of-the-art coding performance achieving 72.7% on SWE-bench with enhanced instruction following
- Superior codebase navigation and multi-file editing with 65% reduction in shortcut behavior
- Enhanced steerability and precision for complex implementations with improved tool integration
- Balanced performance and efficiency ideal for user-facing assistants and high-volume tasks
Web Search
Search the web for current info.
Extended Thinking
Enhanced reasoning for complex tasks.
Web Fetch
Fetch and read full content from any URL.
Code Execution
Run Python & Bash code in a sandbox.
JSON Mode
Output responses in valid JSON format.
Model Information
Supported Formats
Claude Opus 4.8Anthropic's most capable model released May 2026. Hybrid reasoning model with a 1M token context window, stronger across coding, agentic tasks, and professional work over Opus 4.7. Features adaptive thinking via the effort parameter (defaults to high). Knowledge cutoff January 2026.
Claude Opus 4.7Anthropic's most capable generally available model with a step-change improvement in agentic coding over Claude Opus 4.6. Features 1M token context window, 128K max output, and adaptive thinking. Knowledge cutoff January 2026.
Claude Opus 4.6Anthropic's most intelligent model released February 2026. Features 1M token context window and 128K max output. Built for deep reasoning, long-horizon agentic tasks, and complex multi-step coding workflows. Supports extended thinking and adaptive thinking modes.