Anthropic has unveiled its latest AI model, Claude 4.5, designed to deliver stronger performance for enterprise users. The update focuses on longer coding runs, financial reasoning, and scientific applications, highlighting the company’s push to serve professional and business customers.
In internal testing, Claude 4.5 successfully created a web app from scratch and coded continuously for 30 hours. This marks a major leap from earlier models, which managed only seven hours of uninterrupted coding. Such improvements underscore the model’s potential for teams managing complex, multi-step projects.
Chief Product Officer Mike Krieger emphasized that Anthropic is not seeking viral consumer adoption but instead targeting power users and business customers who require reliable and scalable AI support. This approach sets the company apart in an increasingly competitive AI market.
Claude 4.5 also showed significant progress in computer usage tasks, achieving a 60% score on an operating-system benchmark compared to 40% for its predecessor. The model is built to handle finance and scientific reasoning more effectively, strengthening its appeal to regulated industries and research-focused organizations.
Microsoft has also announced plans to integrate Anthropic’s models into its Microsoft 365 Copilot. New features like “Agent Mode” in Excel and Word, and an “Office Agent” in Copilot chat, will leverage Claude’s improved capabilities. PowerPoint integration is expected soon, broadening the productivity benefits for enterprise teams.
Founded by former OpenAI executives, Anthropic continues to position Claude as a workplace-first AI with robust guardrails for safe use. The company has been actively marketing the model’s data-analysis and coding strengths to regulated sectors that demand precision and security.
By focusing on sustained performance and practical business applications, Claude 4.5 marks a step forward in Anthropic’s strategy to deliver reliable AI solutions tailored to business customers worldwide.
In other news read more about: OpenAI Unveils Operator, AI Agent for Autonomous Online Tasks
