Updates

Anthropic Releases Claude Sonnet 4.6 — Most Capable Sonnet Model Yet

Anthropic has released Claude Sonnet 4.6, a full upgrade across coding, computer use, long-context reasoning, agent planning, and design. Features a 1M token context window in beta and is now the default model for Free and Pro plans.

ClaudeAnthropicCodingAI Agents
※ このページにはアフィリエイトリンクが含まれています。リンク経由でご購入いただくと、運営費の一部として還元されます。

Claude Sonnet 4.6 Overview


Anthropic officially released Claude Sonnet 4.6 on February 17. As the most capable Sonnet model yet, it delivers major improvements across coding, computer use, long-context reasoning, agent planning, knowledge work, and design.


Key Improvements


Coding Leap


  • 70% of Claude Code users prefer it over Sonnet 4.5
  • Preferred over Opus 4.5 in 59% of cases
  • Better at consolidating shared logic instead of duplicating code
  • Significantly less prone to overengineering and laziness

  • Computer Use Evolution


  • Steady gains on OSWorld benchmark
  • Human-level capability in complex spreadsheet navigation and multi-step web forms
  • Major improvement in prompt injection resistance

  • 1M Token Context Window


  • Can hold entire codebases, lengthy contracts, or dozens of research papers in a single request
  • Developed unique investment strategy in Vending-Bench Arena, outperforming competitors

  • Pricing


    Same as Sonnet 4.5 at $3/$15 per million tokens. Available as the default model for Free and Pro plans.

    AI Newsletter

    Get the latest AI tools and news delivered daily

    Related Articles