Replit, the AI coding platform popular among startup founders and developers, announced it is no longer working with Anthropic, CEO Amjad Masad said Thursday at TechCrunch Disrupt. The company made OpenAI’s ChatGPT its default AI model in August, shifting away from Anthropic’s Claude models, which it previously relied on to power its platform, according to techcrunch.com.
Masad explained that the decision to move away from Anthropic was driven by a combination of cost, partnership dynamics, and performance metrics. Replit uses internal and public benchmarks to evaluate AI models, and ChatGPT consistently outperformed Anthropic’s Sol and Astra models in building vibe-coded applications from scratch. Despite the shift, Anthropic’s models remain accessible to customers who choose to use them, but most users stick with the default ChatGPT, reducing Replit’s overall spend with Anthropic, techcrunch.com reported.
This change reflects broader trends in the AI development space, where companies prioritize cost efficiency and model performance. Replit’s move to OpenAI aligns it with one of the largest AI providers, potentially enhancing user experience and operational scalability. The decision highlights the competitive landscape among AI model providers, with OpenAI’s ChatGPT gaining preference over Anthropic’s offerings in this context, according to techcrunch.com.
Replit’s CEO made these remarks on October 15 during TechCrunch Disrupt, confirming that while Anthropic is no longer the default, its models remain available on the platform. The company’s partnership with OpenAI began in August, marking a strategic pivot in its AI model integration.