The model can make plans, use tools, operate in Claude Code, and adjust behavior through effort levels. Anthropic reports improvements over Sonnet 4.6 in agentic search, computer use, coding, and knowledge work, alongside safety evaluations that found fewer undesirable behaviors overall and weaker cybersecurity capability than its current Opus models.
Claude Sonnet 5 is useful for developers who need a capable general-purpose model with a balance of quality and cost. It is available across Claude plans and through the Claude Platform API, with introductory pricing of $2 per million input tokens and $10 per million output tokens through August 31, 2026, before moving to $3 and $15.
