September 30, 2025Published by Sean Kim on September 30, 2025Categories AI Tools & ServicesClaude Sonnet 4.5 Release: 77.2% SWE-bench Score and 30-Hour Autonomous Agents — What ChangedAnthropic just dropped Claude Sonnet 4.5, and the numbers speak for themselves: 77.2% on SWE-bench Verified, 61.4% on OSWorld, and agents that can stay focused for […]