Anthropic launches Claude Sonnet 5, positioned as the company’s most agentic mid-tier model and a lower-cost alternative to its flagship offerings. Multiple outlets report that Sonnet 5 delivers agent-focused capabilities designed to “act” and run more autonomously rather than only answer prompts. Anthropic says Sonnet 5 can plan, use tools such as browsers and terminals, and complete complex tasks that earlier Sonnet versions could not. The model also includes self-checking of its output without being prompted.
In performance terms, outlets state that Sonnet 5 narrows the gap with the Opus line, with claims that it reaches close-to-Opus 4.8 performance on many tasks. Several reports also cite improvements over Sonnet 4.6 in areas such as reasoning, tool use, coding, and knowledge work.
On safety and behavior, Anthropic says Sonnet 5 shows lower rates of hallucinations and “sycophancy,” and it refuses malicious requests more effectively. Sonnet 5 is available starting June 30, 2026 across all Anthropic plans, including as the default model for Free and Pro plans. Pricing is reported as $2 per million input tokens and $10 per million output tokens through August 31, then increasing to $3 and $15.
