Anthropic has released Claude Sonnet 4.6, its most capable Sonnet model to date, featuring a 1M token context window and major improvements across coding, computer use, and reasoning tasks. The model is now the default for Free and Pro users on claude.ai and Claude Cowork, maintaining the same pricing as its predecessor at $3/$15 per million tokens.

Sonnet 4.6 delivers performance that previously required Opus-class models, with early developers preferring it to both Sonnet 4.5 and November's Claude Opus 4.5. The model shows significant gains in computer use capabilities, approaching human-level performance on tasks like navigating complex spreadsheets and multi-step web forms. Safety evaluations found it to be as safe as or safer than previous models, with researchers noting its "warm, honest, prosocial" character.

Benchmarks show Sonnet 4.6 matching Opus 4.6 on document comprehension tasks and outperforming its predecessor on long-horizon planning, with users reporting fewer hallucinations and better instruction following. The model's ability to reason effectively across its extended context window enables sophisticated strategies in complex tasks, such as investing capacity early before pivoting to profitability in simulated business scenarios.