Replit, the cloud-based software development platform popular with startup founders and developers, has ended its partnership with AI lab Anthropic, according to CEO and co-founder Amjad Masad. Speaking on stage at TechCrunch Disrupt on Thursday, Masad said the decision came down to several factors, including cost.
Replit announced a partnership with OpenAI in August, but had given no previous indication that it had parted ways with Anthropic. Under the earlier partnership, Replit used Anthropic's Claude models to run its platform, an arrangement reported to have helped the company grow revenue significantly.
"I think ChatGPT is better now," Masad said, adding that the company uses an internal benchmark to assess performance. "We publish a public benchmark as well, and consistently, Sol and Astra perform better on building 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."
However, Masad's comments were rejected by a person familiar with the Anthropic-Replit partnership, who asked not to be named and said the partnership was still in place.
Public evidence suggests Replit was using Anthropic products as recently as a few weeks ago. A late September blog post cites Anthropic's Claude Fable 5 and Claude Fable 5.1 as models running in Replit Agent's production environment, and around the same time a Replit product employee posted on X about using Anthropic's flagship model Opus 5.5. TechCrunch has reached out to Replit for clarification.
Masad also shared details of the company's growth at the conference, saying enterprise contracts made up about 2% of revenue last year but around 15% this year, with more than 50% of revenue now coming from B2B customers.
Discussing AI safety, Masad said agents need a secure environment to work in: "In the same way that us humans need a certain environment to work ... agents need a habitat for them to exist and do good work. So that habitat has good orchestration, good error recovery, good safety and security and sandboxing, good networking, good identity and authorization."