Replit is no longer working with Anthropic, CEO says

Replit, the popular AI coding platform, has decided to terminate its partnership with Anthropic, according to the company's CEO and co-founder, Amjad Masad. During a presentation at TechCrunch Disrupt, Masad cited cost, partnership, and performance as the primary reasons behind this decision. Replit had previously announced a partnership with OpenAI in August, but there was no indication it was ending the collaboration with Anthropic.

Masad highlighted the superior performance of Replit's internal benchmark, stating that both Sol and Astra models outperformed Anthropic's models in building applications from scratch. Despite the termination of the partnership, Replit users can still access Anthropic models on the platform, with OpenAI's ChatGPT now set as the default option.

This shift is expected to significantly cut down Replit's spending with Anthropic, as most customers utilize the default model. Prior to the deal with OpenAI, Replit relied on Anthropic's Claude models to run its platform, which helped the company achieve significant revenue growth. Masad affirmed that Replit maintains a model-agnostic approach, allowing users to choose between Anthropic's Claude or other supported models.

The CEO reiterated that Replit remains committed to Anthropic, as evidenced by the company's website, blog posts, and social media posts, which highlighted the use of Anthropic's Claude Fable 5 and Opus 5.5 models as recently as late September. However, insiders familiar with the partnership dispute this claim. Replit's CEO also acknowledged the importance of strong AI safety measures, emphasizing the need for a secure habitat to ensure safe deployment of AI agents.

He expects more than half of Replit's revenue to come from enterprise and government customers by 2027.