- Published on
Replit is no longer working with Anthropic, CEO says
- Authors
- Name
- DeepHorus
Replit, the cloud-based AI coding platform, has ended its work with Anthropic, CEO Amjad Masad said Thursday at TechCrunch Disrupt. The split was driven by cost, partnership, and performance, he said. Replit had used Anthropic’s Claude models to power its platform, a relationship that reportedly helped revenue grow. Masad said OpenAI’s ChatGPT is better now.
The move follows Replit’s August announcement of a partnership with OpenAI. Until Masad’s comments, there was no public indication that the Anthropic relationship had ended. Replit sells a browser-based software development environment where users build and deploy apps directly. Its agent products rely on large language models, making model partnerships central to performance and cost.
Masad said Replit uses internal and public benchmarks. He claimed Sol and Astra consistently perform better at building “vibe-coded applications from scratch.” “I think ChatGPT is better now,” Masad said, adding that cost and partnership also mattered. He called OpenAI a better partner “for now.”
Not everyone agrees. Insiders familiar with the partnership rejected Masad’s account, and one person said it remained in place. Public evidence shows Replit using Anthropic products recently: a late September blog post cited Claude Fable 5 and Claude Fable 5.1 in Replit Agent’s production environment.
Masad also shared growth figures at Disrupt. Enterprise contracts were about 2% of revenue last year, he said, but around 15% this year. More than 50% of revenue now comes from B2B customers. Replit is not yet sharing ARR numbers.
On AI safety, Masad compared agents to humans needing a good habitat. He said orchestration, error recovery, security, sandboxing, networking, identity, and authorization are essential. Once solved, he argued, AI can be deployed safely. TechCrunch said it would update the story if it heard back.
