Replit Is No Longer Working With Anthropic, CEO Says
Amjad Masad cites cost and model performance as the coding platform tilts toward OpenAI, though an insider says the relationship remains active.
Software development platform Replit has ended its working relationship with AI research lab Anthropic, co-founder and Chief Executive Amjad Masad stated onstage at the TechCrunch Disrupt conference, as reported by TechCrunch AI (https://techcrunch.com/2026/10/16/replit-is-no-longer-working-with-anthropic-ceo-says/).
Masad pointed to model capabilities, financial terms, and partnership alignment as the key drivers behind the change. According to Masad, OpenAI's models currently outperform Anthropic's on internal and public evaluation benchmarks used by Replit to gauge performance when generating applications from scratch.
The move follows a formal partnership that Replit established with OpenAI in August. Replit, which provides a browser-based environment for building and deploying software, had previously relied on Anthropic's Claude models to power its automated coding agents—an integration credited with driving substantial revenue growth for the company.
Masad's onstage comments, however, drew immediate pushback from an insider familiar with the arrangement, who told TechCrunch on condition of anonymity that the partnership between Replit and Anthropic remains in place. Production records show recent Anthropic usage, including a September blog post identifying Claude Fable 5 and Claude Fable 5.1 inside Replit Agent's live environment.
Beyond provider partnerships, Masad outlined shifts in Replit's revenue composition, reporting that enterprise deals have expanded from roughly 2 percent of revenue last year to approximately 15 percent this year. Masad declined to disclose specific annual recurring revenue figures, but added that business-to-business contracts now account for more than 50 percent of the startup's total revenue.
Addressing operational security, Masad emphasized that autonomous coding agents require robust infrastructure to run reliably in production. He noted that safe agent deployment depends on managed digital habitats equipped with sandboxing, identity authorization, network controls, orchestration, and systematic error recovery.
Sources
Written by
The Company Wire
Inside the companies building what’s next. Reporting on startups, technology, funding and the people shaping them.
