Replit Ends Partnership with Anthropic, CEO Says

Replit has ended its partnership with Anthropic, according to CEO Amjad Masad. He confirmed the company is no longer working with the Claude maker, citing several reasons, including cost. The move became public after questions about Replit's AI model strategy, following its August partnership with OpenAI. Until now, there had been no clear sign that the Anthropic relationship had closed, making the confirmation a notable shift in a fast-moving market. The decision puts Replit more firmly in OpenAI's orbit as it builds and markets AI coding tools. For developers, the most immediate questions are what models power Replit's assistant and whether the change affects response quality, latency, or subscription pricing. Cost is a major factor for any platform that serves many users. Running large language models at scale can be expensive, and companies often renegotiate or switch providers to control margins. Replit may also be seeking tighter integration with OpenAI's APIs and product roadmap. The split underscores a broader competition among AI coding assistants. Startups and established platforms must choose among model providers such as OpenAI, Anthropic, Google, and open-weight alternatives. Each offers different strengths in code generation, debugging, long-context reasoning, and tool use. Those choices can shape user experience and the economics of a product. Anthropic's Claude models have been popular for coding tasks, so Replit's departure may surprise some in the developer community. However, vendor relationships in AI are rarely permanent. Pricing, performance, reliability, and strategic alignment all change quickly. For Replit, the challenge is to show that the switch does not degrade its platform. If OpenAI's models handle coding tasks well and cost less, the partnership could strengthen Replit's position. If users notice weaker results, the company may face criticism. The episode also highlights how dependent AI products are on upstream model makers. As competition intensifies, more coding platforms may reassess their suppliers, making model partnerships a key part of product strategy rather than a back-end detail.