Anthropic has released Claude Sonnet 4.6, its most capable Sonnet-class model to date, featuring major upgrades in coding, computer use, long-context reasoning with a 1M token context window (in beta), agent planning, and design capabilities. The model achieves strong benchmark results, including 79.6% on SWE-bench Verified and 72.5% on OSWorld-Verified, while maintaining the same API pricing and becoming the default for free and Pro users on claude.ai. It also shows improved resistance to prompt injection attacks, making it more secure for agentic and cybersecurity-relevant workflows.

This is Claude Sonnet 4.6: our most capable Sonnet model yet.

It's a full upgrade across coding, computer use, long-context reasoning, agent planning, knowledge work, and design.

It also features a 1M token context window in beta. pic.twitter.com/TDId3XUSRs

— Claude (@claudeai) February 17, 2026