Anthropic unveiled its Claude 4.5 artificial intelligence model on Monday, saying the newest version can code for longer uninterrupted periods and better handle financial and scientific tasks, in a move by the startup toward enterprise AI.
The AI startup backed by Alphabet and Amazon.com competes with its rivals to develop models that can reliably operate software and complete multi-step tasks, which is key for AI agents that can perform tasks for humans.
The Sonnet 4.5 model created a web application from scratch in internal testing, and one customer autonomously maintained the AI chatbot code for 30 hours, compared to seven hours for Anthropic’s previous Claude Opus 4 for a different user, said product director Mike Krieger.
Anthropic is focusing on advanced users and enterprise customers rather than viral consumption, he said.
Claude 4.5 is stronger in finance and scientific reasoning and better at using computers, scoring around 60% on a test that analyzes operating system proficiency, compared to approximately 40% for previous models, the company said.
Meanwhile, Microsoft said Monday that it would add new Microsoft 365 Copilot features powered by Anthropic models, including “Agent Mode” in Excel and Word, and an “Office Agent” in Copilot chat.
Last week, Microsoft said it would bring Anthropic’s models to Microsoft 365 Copilot to diversify away from its longtime partner OpenAI.
Anthropic, founded by former OpenAI executives, has positioned Claude for use in the workplace with security measures that they say reduce risky outcomes.
The company has been promoting Claude’s coding and data analysis skills to regulated industries and teams that want the models to work across multiple software tools.
Source: Yahoo!Finanzas
