Replit CEO and co-founder Amjad Masad said on Thursday that the AI coding platform is no longer working with Anthropic, a break that caught many observers off guard given the two companies had been closely aligned.
Speaking onstage at TechCrunch Disrupt, Masad cited cost, performance, and the strength of Replit's new relationship with OpenAI. "I think ChatGPT is better now," he said, noting that the company's internal and public benchmarks show its Sol and Astra models perform better than Claude at building "vibe-coded" applications from scratch.
Replit announced a partnership with OpenAI in August, and before that deal the company had leaned heavily on Anthropic's Claude models to run its platform. Masad noted that Replit had been one of Anthropic's top five customers.
The shift may not be as complete as it sounds. Masad later clarified that customers can still access Anthropic models on the platform, but OpenAI's ChatGPT is now the default. "Replit remains model agnostic," he said.
The announcement drew skepticism. Insiders familiar with the partnership told TechCrunch the Anthropic relationship was still in place, and a September Replit blog post still listed Claude in Replit Agent's production environment.