Replit Switches Default to OpenAI's ChatGPT, Says CEO, Though Anthropic Models Remain Available

Replit, the browser-based AI coding platform popular with startup founders and developers, has stopped working with Anthropic, CEO and co-founder Amjad Masad said Thursday onstage at TechCrunch Disrupt.

Masad cited several reasons for the shift, including cost. "I think ChatGPT is better now," he said, pointing to the company's internal benchmark and a public one it also publishes. He said the decision came down to "cost, partnership, and performance," adding that the company felt OpenAI was "a better partner for now."

The move follows Replit's August partnership announcement with OpenAI. Masad later clarified that customers can still access Anthropic models on the platform, but OpenAI's ChatGPT is now the default. Because most customers stick with the default rather than choosing a model themselves, the change could significantly cut Replit's spending with Anthropic. Before the OpenAI deal, Replit ran its platform on Anthropic's Claude models, a partnership credited with helping grow its revenue; Masad said Replit had been, until recently, a top-five Anthropic customer. "Replit remains model agnostic," he told TechCrunch over the weekend, describing OpenAI as its "preferred partner for powering Replit Agent" based on performance, economics, and partnership strength.

The account was disputed, however. One person familiar with the matter, speaking anonymously, said the Anthropic partnership was still in place. Replit's own website, blog posts, and social media promoted Anthropic products as recently as a few weeks ago, including a late-September blog post naming Claude models running in Replit Agent's production environment.

On growth, Masad said enterprise was about 2% of revenue last year and roughly 15% this year, with more than half of revenue now B2B. Backstage, he told TechCrunch he expects more than half of revenue to come from enterprise and government customers in 2027. He also stressed the need for strong AI safety measures, comparing the requirements for AI agents to a human work environment: "agents need a habitat for them to exist and do good work." That habitat, he said, includes orchestration, error recovery, security and sandboxing, networking, and identity and authorization.

Source: TechCrunch