Anthropic has rolled out an upgraded version of its AI model called Claude Sonnet 4.6. This update focuses on making the AI better at coding, following instructions, and handling complex reasoning tasks. Launched on February 17, the new version is a significant step forward in AI performance for a range of technical and knowledge-based tasks.

Enhanced Capabilities in Coding and Reasoning

Claude Sonnet 4.6 improves on previous versions by delivering more consistent results when coding and working with instructions. Developers who have early access report that the model now performs tasks that once required much more advanced AI systems. For example, it can navigate complicated spreadsheets or fill out multi-step web forms with human-like accuracy. The model can even manage multiple browser tabs, pulling together information from different sources seamlessly.

This upgrade means that tasks traditionally done by high-end AI models are now more accessible. The model’s ability to process longer contexts—up to one million tokens in beta—helps it maintain understanding over extended conversations or complex projects. These improvements make the AI more useful for real-world office work, reducing the need for expensive, specialized models.

New Features and Platform Availability

Claude Sonnet 4.6 is now available across all of Anthropic’s platforms, including Claude Cowork, Claude Code, the API, and major cloud services. Developers can access the model through the Claude API by using the identifier claude-sonnet-4-6. The update also introduces features like adaptive thinking and extended reasoning, which help the AI handle more complex problem-solving scenarios.

An exciting addition is the beta version of context compaction. This feature automatically summarizes older parts of a conversation or project, making room for new information without losing important details. This allows the AI to work effectively even as conversations grow longer, improving overall efficiency and response quality.

Anthropic has also upgraded its API, making tools like web search and fetch more powerful. Now, these tools can automatically generate and run code to filter search results, ensuring only relevant content is used. Other new capabilities include improved code execution, memory management, and programmatic tool calling, all of which help developers build smarter applications with the AI.

While Claude Sonnet 4.6 still doesn’t match the skills of top human experts in using computers, it has made significant progress. The rate of improvement suggests that AI models will continue to grow more capable and useful for a wide range of work tasks, from simple data entry to complex decision-making. Overall, this update marks a big step forward in making AI a practical tool for everyday professional use.

Inspired by

• https://www.infoworld.com/article/4133578/claude-sonnet-4-6-improves-coding-skills.html

Sources

• anthropic.com
• claude.com
• claude.com
• claude.com
• claude.com