Replit, a cloud-based development platform that lets users build and deploy apps from their browser, is no longer working with Anthropic, Replit CEO and co-founder Amjad Masad said onstage at TechCrunch Disrupt on Thursday.
Masad attributed the decision to multiple factors including cost, performance and partnership fit. He said Replit uses internal benchmarks to evaluate model performance and also maintains a public benchmark; according to him, the Sol and Astra models have consistently performed better for building so-called "vibe-coded" applications from scratch. "It was a decision based on both cost, partnership, and performance. And for all these reasons, we just felt like they were a better partner for now," he said.
Masad also noted that Replit announced a partnership with OpenAI in August, but there had been no prior explicit indication that the company had stopped working with Anthropic.
Conflicting signals about the Anthropic relationship
Masad’s statement was met with pushback from people familiar with the Anthropic–Replit collaboration. One source who asked not to be named said the partnership remained in place. Additionally, there is evidence that Replit used Anthropic products as recently as September: a September blog post cites Anthropic’s Claude Fable 5 and Claude Fable 5.1 as models running in Replit Agent’s production environment.
TechCrunch reached back out to Replit for clarification; the story will be updated if the company responds.
Why the change matters
Replit’s prior use of Anthropic models was reported to have helped the company grow its revenue significantly. If the partnership has indeed ended, it could affect Replit’s cost structure, product architecture and choices about which large models to integrate into the platform.
Business trends at Replit
Onstage, Masad also described how Replit’s revenue mix is shifting toward enterprise customers. "We’re not sharing ARR numbers just yet, but in terms of growth, like, I would say, enterprise was like 2% of revenue last year," he said. "This year, I would say like 15% of our revenue is enterprise contracts, and more than 50% of our revenue is B2B."
AI safety and operational needs
Masad emphasized the need for robust AI safety and operational infrastructure. He compared agents to humans who need a proper environment to work in — good orchestration, error recovery, safety and sandboxing, networking, identity and authorization. "Once you figure all of that out, then you can deploy AI safely," he said.
Summary
Replit’s CEO says the company has moved away from Anthropic models for reasons of cost, performance and partnership fit and has leaned toward other providers such as OpenAI. Nevertheless, recent references to Anthropic models in Replit’s own documentation and comments from insiders suggest the exact status of the relationship may require further clarification.
