Masad explained that this decision was driven by multiple factors, most notably cost. Although Replit announced a partnership with OpenAI in August, there was never any public indication that it had ceased working with Anthropic.
“I think ChatGPT is better now,” Masad stated, noting that the company utilizes an internal benchmark to evaluate key metrics such as performance. “We also publish a public benchmark, and consistently, Sol and Astra perform better on building vibe-coded applications from scratch. This decision was based on cost, partnership dynamics, and performance. For all these reasons, we simply felt they were the better partner for now.”
Replit operates a cloud-based software development platform that allows users to build and deploy applications directly within their browser. Under its previous agreement with Anthropic, the company utilized the AI lab’s Claude models to power its platform; this collaboration was reportedly instrumental in driving significant revenue growth for the business.
However, Masad’s remarks were met with skepticism from insiders familiar with the relationship between Anthropic and Replit. One anonymous source familiar with the matter claimed that the partnership was still very much active.
Evidence suggests that Replit was still utilizing Anthropic products as recently as September. For example, a blog post from that month references Anthropic’s Claude Fable 5 and Claude Fable 5.1 models running within Replit Agent’s production environment.
TechCrunch reached out to Replit for further clarification and will update this report should we receive a response.
At Disrupt, Masad offered insights into the company’s expansion, highlighting that enterprise clients are increasingly becoming a larger portion of the revenue stream.
“We aren't sharing ARR numbers just yet, but in terms of growth, I would say enterprise accounted for about 2% of revenue last year,” Masad stated. “This year, enterprise contracts represent 15% of our revenue, while more than 50% of our total revenue is B2B.”
Masad also emphasized the critical importance of robust AI safety measures.
“Just as humans require a specific environment to work—such as a nice office with good food, air conditioning, and computers—agents require a habitat to exist and perform effectively,” he explained. “That habitat must feature good orchestration, error recovery, safety and security, sandboxing, networking, and identity and authorization. Once these elements are established, one can safely deploy AI.”
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.