Anthropic unveils latest AI model, aiming to extend its lead in coding intelligence (www.businessinsider.com)

Anthropic has released Claude Sonnet 4.5, its latest generative AI model that the company bills as the “world’s best” coding system. Arriving just four months after Sonnet 4, Sonnet 4.5 claims state-of-the-art performance on SWE-Bench Verified and is optimized for producing production-ready code, improved refactoring judgment, and higher reliability. Anthropic also highlights major business and operational gains: Claude Code is driving more than $500M in run-rate revenue with usage up over 10x in three months, and the new model can run autonomously for up to 30 hours—more than four times the endurance of its prior Opus 4 model—enabling longer-lived autonomous agents and end-to-end automation workflows. Technically, Sonnet 4.5 is pitched not just as a coding assistant but as a platform for autonomous computer use, faster vulnerability detection and remediation in cybersecurity, and superior performance on finance tasks compared with Anthropic’s Opus 4.1. To help developers deploy these capabilities, Anthropic launched the Claude Agent SDK and expanded tooling—native VS Code extension, enhanced terminal workflows, checkpoints for instant rollback, plus virtual machines, memory and context management—so teams can build customized, context-aware agents and production applications. The release is positioned to solidify Anthropic’s lead against rivals like Google’s Gemini, OpenAI’s GPT-5, and xAI’s Grok 4 in the lucrative code-generation market.