Anthropic Launches Claude Sonnet 4.5

Anthropic has launched Claude Sonnet 4.5, its most powerful and aligned model to date, featuring major advances in coding, reasoning, and computer use. It leads benchmarks like SWE-bench Verified (82%) and OSWorld (61.4%), sustaining 30+ hours on complex tasks. New features include checkpoints in Claude Code, a VS Code extension; enhanced context editing; memory tools; code execution; and file creation across apps. Developers gain access to the Claude Agent SDK, which offers the same infrastructure that Anthropic uses for agent design. Safety is enhanced with AI Safety Level 3 protections and reduced instances of misaligned behaviors, including deception and sycophancy. Early users report improved performance in finance, law, medicine, and STEM fields. Pricing remains $3/$15 per million tokens, matching Sonnet 4.

Anthropic, News – Introducing Claude Sonnet 4.5 – 29th September 2025