Anthropic has launched Claude Sonnet 4.5, which it dubs as the "best coding model in the world" and "the strongest model for building complex agents".

The model also shows improved capabilities on a broad range of evaluations including reasoning and math, it said in a blogpost.

On OSWorld, a benchmark that tests AI models on real-world computer tasks, Sonnet 4.5 leads at 61.4%, it noted. Sonnet 4 held the lead at 42.2%, just four months ago, it added.

Claude Sonnet 4.5 is being released under Anthropic's AI Safety Level 3 (ASL-3) protections, it said. 

"These safeguards include filters called classifiers that aim to detect potentially dangerous inputs and outputs—in particular those related to chemical, biological, radiological, and nuclear (CBRN) weapons," it added.

The blogpost also quoted Michael Truell, CEO of Cursor as saying, "We're seeing state-of-the-art coding performance from Claude Sonnet 4.5, with significant improvements on longer horizon tasks."