Masad explained that several factors drove the decision, with cost being one of them. Replit had announced a partnership with OpenAI back in August, yet nothing suggested it was ending its work with Anthropic. "I think ChatGPT is better now," Masad remarked, noting that the company relies on an internal benchmark to evaluate things like 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."

Still, it appears Anthropic hasn't been entirely removed from Replit, which offers a cloud-based software development platform allowing users to build and deploy apps right in their browser. Following Disrupt, Masad made clear that customers can still access Anthropic models on the platform, though OpenAI's ChatGPT is now the default option. This change could substantially cut Replit's spending with Anthropic, given that most customers stick with the default rather than picking a model on their own. Before the OpenAI deal, Replit had depended on Anthropic's Claude models to power its platform. That partnership was reportedly instrumental in boosting the company's revenue significantly, and Masad noted on stage that Replit had been, until recently, one of Anthropic's top-five customers. "Our users can choose Anthropic's Claude or other supported models…OpenAI is now our preferred partner for powering Replit Agent, based on performance, economics, and the strength of our partnership."

However, insiders familiar with the Anthropic-Replit partnership pushed back on Masad's onstage remarks. One person with knowledge of the situation, who requested anonymity, said the partnership remained intact. Evidence on the company's website, blog posts, and social media shows that Replit was still promoting Anthropic products as recently as a few weeks ago. For example, a late September blog post references Anthropic's Claude Fable 5 and Claude Fable 5.1 as models running in Replit Agent's production environment. Around the same period, a Replit product employee posted on X about using Anthropic's flagship model Opus 5.5 and touted its capabilities.

Masad also shared growth insights while onstage at Disrupt, pointing out that enterprise is beginning to account for a bigger share of revenue. "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," Masad said. "This year, I would say like 15% of our revenue is enterprise contracts, and more than 50% of our revenue is B2B."

Masad also weighed in on the need for solid AI safety measures. "In the same way that us humans need a certain environment to work - you need a nice office, maybe nice good food in the office, air-conditioning, computer - agents need a habitat for them to exist and do good work," he said. "So that habitat has good orchestration, good error recovery, good safety and security and sandboxing, good networking, good identity and authorization. And once you figure all of that out, then you can deploy AI safely."

This article was originally published on October 16 at 2:22 pm PT. It has been updated to include more context around Replit's partnerships with Anthropic and OpenAI.