Replit Drops Anthropic for OpenAI: What It Means for AI Coding

Replit, a company that lets you build apps right in your web browser using AI, is no longer working with Anthropic, the maker of the Claude AI models. CEO Amjad Masad announced the change at a tech conference, saying Replit now uses OpenAI's models instead. He cited three reasons: lower cost, better performance on Replit's internal tests, and a stronger partnership with OpenAI. Replit had partnered with OpenAI in August, but this is the first confirmation that it's no longer using Anthropic.

Why should you care? Replit is used by many startups and even non-coders to create software quickly. The AI models behind it affect how well and how fast your apps get built. If OpenAI's models are indeed cheaper and better, Replit users could see lower prices or improved results. However, some people close to Anthropic say the partnership is still active, and Replit's own blog posts from late September show they were still using Anthropic's models then. So there's confusion about what's really happening.

Masad also shared that Replit's business is growing, with enterprise customers now making up 15% of revenue, up from 2% last year. He emphasized the need for good AI safety measures, comparing AI agents to humans who need a good work environment. For everyday users, this means Replit is focusing on making AI coding safer and more reliable, which could lead to better tools for everyone.

The takeaway: if you use Replit or are thinking about it, the AI powering it may be shifting, which could affect your experience. Keep an eye on updates, as the situation is still developing.