Replit’s CEO says the company has ended its partnership with Anthropic

Meta description: Replit CEO Amjad Masad said the company no longer works with Anthropic, but sources and recent posts suggest the partnership may remain unresolved.

Tags: Replit, Anthropic, AI coding, OpenAI, TechCrunch Disrupt


Replit, the AI coding platform favored by startup founders and developers, became a point of confusion after CEO and co-founder Amjad Masad said Thursday onstage at TechCrunch Disrupt that the company is no longer working with Anthropic. The remark mattered because Replit had publicly partnered with Anthropic's Claude models, and model choice shapes both the developer experience and the company's cost structure.

Background: Replit's model stack and earlier partnerships

Replit sells a cloud-based software development platform where users can develop and deploy apps directly from inside their browser. Under its partnership with Anthropic, Replit used the AI lab's Claude models to run its platform. That arrangement was said to have helped the company grow its revenue significantly.

In August, Replit announced a partnership with OpenAI. At the time, however, there was never any public indication that it was no longer working with Anthropic. The latest statement from Masad therefore arrived without a prior announcement of a split.

Development: Masad cites cost, performance and partnership

Masad said the decision was made for a number of reasons, including cost. "I think ChatGPT is better now," he said, adding that Replit uses an internal benchmark to help assess metrics like performance. He also said Replit publishes a public benchmark and pointed to its results, saying Sol and Astra consistently perform better on building vibe-coded applications from scratch. "It was a decision based on both cost, partnership, and performance," Masad said. "And for all these reasons, we just felt like they were a better partner for now."

That rationale places the change inside a broader pattern for AI coding platforms: model providers are evaluated not only on raw capability, but also on price, integration ease, benchmark results for a specific product workflow, and the terms of the partnership itself.

Latest: Anonymous accounts and recent posts conflict with Masad

Masad's comments were rejected by insiders familiar with the partnership between Anthropic and Replit. One person familiar with the situation, who asked not to be named, said the partnership was still in place.

There is also evidence on Replit's website, blog posts, and social media posts that the company was using Anthropic products as recently as a few weeks ago. A blog post in late September cites Anthropic's Claude Fable 5 and Claude Fable 5.1 as models running in Replit Agent's production environment. Around the same time, one of Replit's product employees posted on X about using Anthropic's flagship model Opus 5.5 and promoting its capabilities.

TechCrunch reached back out to Replit for clarification and said it would update the story if it heard back. The outlet also updated the article to include more context on Replit's use of Anthropic models.

Context: Enterprise revenue and AI safety takeaways

Masad used the same session to describe Replit's growth. He said the company was not sharing annual recurring revenue figures just yet, but noted that enterprise was about 2% of revenue last year. This year, he said, about 15% of revenue is enterprise contracts, and more than 50% of revenue is business-to-business.

He also discussed the need for good AI safety measures. He compared agents to humans who need a certain environment to work, including a nice office, maybe good food in the office, air-conditioning, and a computer, saying agents need a habitat to exist and do good work. He said that habitat should have good orchestration, good error recovery, good safety and security and sandboxing, good networking, and good identity and authorization. Once those elements are figured out, he said, AI can be deployed safely.

The current stage is unresolved. Masad presented the Anthropic relationship as a past partnership, while recent Replit materials and anonymous accounts suggest the company may still be using Anthropic models. Until Replit or Anthropic offers a clearer statement, the split remains a claim rather than a fully documented transition.