Anthropic has released Claude Sonnet 5, a new version of its mid-tier AI model designed to handle more autonomous work without pushing users into its more expensive Opus models.
The company introduced the model on June 30, 2026, saying it is now available across all Claude plans and has become the default model for Free and Pro users. Max, Team and Enterprise customers can also access it, along with developers using Claude Code and the Claude Platform.
Anthropic described the release as a step toward more practical AI agents, saying: “Claude Sonnet 5 is built to be the most agentic Sonnet model yet.” The company said the model can make plans, use tools such as browsers and terminals, and run with a level of autonomy that recently required larger models.
The pitch is partly about price. Sonnet 5 launches at $2 per million input tokens and $10 per million output tokens through August 31, 2026. After that, Anthropic said pricing will rise to $3 per million input tokens and $15 per million output tokens.
Anthropic said Sonnet 5 performs close to Claude Opus 4.8 on some tasks while costing less, especially for coding, reasoning, tool use and knowledge work. Developers can call the model through the Claude API using the model name claude-sonnet-5.
The company also tried to draw a line around safety. Anthropic said its pre-deployment assessments found Sonnet 5 showed a lower rate of undesirable behavior than Sonnet 4.6 and was safer in agentic settings, though it also said the model showed somewhat higher rates of misaligned behavior than Opus 4.8 and Claude Mythos Preview.
Sonnet 5 will ship with cyber safeguards turned on by default. Anthropic said the model was never deliberately trained for cybersecurity tasks and performed worse than its current Opus models on tests involving dangerous cyber capabilities.
