Anthropic released Claude Sonnet 5 on June 30, positioning it as the most agentic model in the Sonnet line so far. It's a step up from Sonnet 4.6 on reasoning, tool use, coding, and general knowledge work, and Anthropic says it narrows the performance gap to the larger Opus 4.8 model while running at a fraction of the cost.
Safety testing found Sonnet 5 has a lower rate of hallucination, sycophancy, and other undesirable behaviors than its predecessor, though it still trails Opus-class models specifically on cybersecurity task performance, by design, since Anthropic didn't train it for that.
It ships today as the default model for Free and Pro users, with availability on Max, Team, Enterprise, Claude Code, and the API. Pricing is $2 per million input tokens and $10 per million output tokens through August 31, rising to $3/$15 after that.
[link] [comments]