Replit, the popular AI coding platform that has captured the attention of startup founders, software engineers, and developers worldwide, is no longer working with AI lab Anthropic. The surprising revelation was delivered by Amjad Masad, CEO and co-founder of Replit, while speaking onstage at TechCrunch Disrupt.
During his appearance at the prominent technology conference, Masad explained that the decision to part ways with Anthropic was driven by a variety of factors, chief among them being cost considerations and overall platform performance. While Replit had officially announced a major partnership with rival AI giant OpenAI back in August, there had been no prior public indication or formal announcement suggesting that the platform had completely severed its ties with Anthropic, whose Claude models previously powered critical infrastructure for Replit’s software development tools.
"I think ChatGPT is better now," Masad told the audience during his session at TechCrunch Disrupt. He elaborated that Replit continuously evaluates its underlying AI models using a comprehensive internal benchmarking system designed to rigorously assess performance metrics across multiple dimensions.
"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," Masad added.
Replit operates a widely utilized cloud-based software development platform that allows users to write, build, test, and deploy applications directly from inside their web browsers. Under the terms of its previous collaboration with Anthropic, Replit heavily integrated the AI lab’s Claude models to run its core platform features. That high-profile partnership was widely credited by industry observers with helping Replit scale its business operations and grow its overall revenue significantly over the past year.
However, Masad’s public assertion that the collaboration has ended was met with immediate pushback from insiders familiar with the ongoing business relationship between Anthropic and Replit. Speaking on the condition of anonymity, one individual familiar with the matter insisted that the partnership was still very much in place, casting a layer of uncertainty over the exact status of the corporate alliance.
Further complicating the narrative, concrete evidence across Replit’s official web properties, technical documentation, corporate blog posts, and public social media channels indicates that Replit was actively utilizing Anthropic products as recently as a few weeks prior to Masad’s onstage announcement. For example, a detailed blog post published by Replit in late September explicitly cites Anthropic’s Claude Fable 5 and Claude Fable 5.1 models as operating actively within the production environment of Replit Agent, the company’s autonomous coding assistant. Around that exact same timeframe, a Replit product employee took to the social media platform X to post about utilizing Anthropic’s flagship model, Opus 5.5, while actively promoting its advanced engineering capabilities to developers.
In light of the conflicting accounts between Masad’s public statements and the recent technical documentation, TechCrunch reached back out to Replit seeking further clarification regarding the current status of the Anthropic integration.
Beyond the dramatic revelations surrounding its AI model providers, Masad used his time onstage at TechCrunch Disrupt to share broader insights into Replit’s financial growth and shifting business model, highlighting a notable pivot toward enterprise customers as a primary growth driver.
While Masad declined to disclose specific annual recurring revenue (ARR) figures during the panel discussion, he offered a detailed breakdown of how the company’s revenue composition has shifted dramatically over the past year. According to the CEO, enterprise contracts accounted for a mere 2% of the company’s total revenue during the previous year.
"This year, I would say like 15% of our revenue is enterprise contracts, and more than 50% of our revenue is B2B," Masad noted, illustrating a massive surge in commercial adoption as businesses increasingly embrace AI-powered development tools to streamline their engineering workflows.
The conversation at Disrupt also touched on the broader technological landscape of autonomous artificial intelligence, with Masad emphasizing the critical industry-wide necessity for robust AI safety measures and secure operational environments as coding agents become more autonomous.
"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," Masad explained to the audience.
He emphasized that creating a proper digital habitat for autonomous agents requires sophisticated infrastructure, including reliable orchestration layers, robust error recovery mechanisms, stringent safety, security, and sandboxing protocols, advanced networking capabilities, and proper identity and authorization controls.
"And once you figure all of that out, then you can deploy AI safely," Masad concluded.
Comments (0)