Claude 4.5 Released: The Most Powerful AI Programming Model Yet

Anthropic has released Claude Sonnet 4.5, marking a major breakthrough in AI programming capabilities. In the SWE-bench Verified benchmark, Claude Sonnet 4.5 ranked first in the industry, capable of working continuously for over 30 hours and generating around 11,000 lines of code in a single run. It also achieved a leading score of 61.4% on the OSWorld benchmark. Additionally, Claude Sonnet 4.5 introduces a new “checkpoint” feature for saving progress at any time, a revamped terminal interface, and a native VS Code plugin.