Anthropic released Claude Sonnet 5 on June 30. It made the model the default for Free and Pro plans and available to Max, Team, and Enterprise users, as well as in Claude Code and through its application programming interface (API).

The company said Sonnet 5 can plan tasks and use tools such as browsers and terminals.

It costs US$2 per million input tokens and US$10 per million output tokens through August 31, 2026, before rising to US$3 and US$15.

Anthropic said Sonnet 5 scored 63.2% on an agentic coding benchmark, compared with 58.1% for Sonnet 4.6 and 69.2% for Opus 4.8, but its public release post did not name the benchmark.

The company also said Sonnet 5 showed lower rates of undesirable behaviors than Sonnet 4.6, including cooperation with misuse and deception.

Anthropic added that the model also had lower rates of hallucination and sycophancy, and was better at refusing malicious requests and resisting prompt-injection attacks.

🔗 Source: Anthropic