What is Claude Sonnet 5?

Claude Sonnet 5 is the most agentic Sonnet model yet from Anthropic, designed to plan, use tools like browsers and terminals, and run autonomously at a level that previously required larger, more expensive models. It delivers top-tier intelligence for coding and everyday professional work, narrowing the performance gap with Opus-class models while maintaining lower pricing. Available across all plans—Free, Pro, Max, Team, and Enterprise—it also powers Claude Code and the Claude Platform, with introductory pricing through August 31, 2026.

Who it's for

• Developers who need a highly capable model for sustained coding, debugging, and multi-step software engineering tasks.
• Knowledge workers looking for an autonomous assistant that can handle complex professional workflows, from research to document creation.
• Teams building agentic automations that require reliable tool use, follow-through, and cost efficiency across messy technical contexts.

Key features

Agentic autonomy

Claude Sonnet 5 can make plans, use browsers and terminals, and run autonomously to complete complex tasks. It finishes multi-step jobs where previous Sonnet models would stop short, and checks its own output without being asked.

Top-tier coding and tool use

The model shows substantial improvements over Sonnet 4.6 in reasoning, tool use, and coding. It handles sustained coding sessions across messy technical contexts, making it a strong execution layer for software engineering work.

Cost-performance flexibility

Sonnet 5 offers a wider range of cost-performance options than its predecessor. At medium effort, it provides substantially improved cost efficiency; at higher effort, it can match Opus 4.8 on some tasks. Introductory pricing runs at $2 per million input tokens and $10 per million output tokens through August 31, 2026.

Improved safety in agentic contexts

Safety assessments show Sonnet 5 has an overall lower rate of undesirable behaviors than Sonnet 4.6, and is generally safer to use in agentic workflows. It also has a much lower ability to perform cybersecurity tasks than current Opus models.