Replit's CEO Said Onstage He Was Done With Anthropic, and People Close to the Deal Say That's Not True

Amjad Masad walked onto the stage at TechCrunch Disrupt on Thursday and announced a breakup. Replit, the coding platform he cofounded and runs, is "no longer working with Anthropic," he told the crowd. Then he added the line that stung: "I think ChatGPT is better now."

It was a remarkable thing to say in public about a supplier that helped build your business. Before Replit signed with OpenAI in August, its platform ran on Anthropic's Claude models, a partnership credited with helping Replit grow revenue significantly. Masad himself said Replit had been, until recently, one of Anthropic's top five customers.

The reasons he gave

Masad framed it as a numbers decision. Replit runs an internal benchmark that tracks things like performance, and he pointed to its public one too. "Consistently, Sol and Astra perform better on building vibe-coded applications from scratch," he said. "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."

Cost is the part a lot of companies will recognize. AI bills have become a line item that finance teams now police, the same pressure that led Accenture to cap what employees can spend on Claude at $29 a month.

The walk-back

By the weekend the story had softened. Masad clarified that Replit customers can still pick Anthropic's models, but OpenAI's ChatGPT is now the default. "Replit remains model agnostic," he said in a statement to TechCrunch. "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."

That sounds like a smaller change than a breakup. In practice it may not be. Most Replit users never touch the model picker and simply run whatever the default is, which means the switch could sharply cut what Replit spends with Anthropic even if Claude stays on the menu.

The other side of the story

People familiar with the Anthropic and Replit relationship pushed back on how Masad described it onstage. One of them, who was not named, said the partnership is still in place. Anthropic has not commented on the record.

The paper trail also complicates the clean narrative. In late September, a Replit blog post described Anthropic models running in Replit Agent's production environment. Around the same time, a Replit product employee was on X promoting Opus, Anthropic's flagship model. Nothing in public suggested the relationship was ending, even after the OpenAI deal was announced. TechCrunch later changed its headline from Replit being done with Anthropic to Replit making OpenAI its default.

Why the default is a business decision now

Masad would not share Replit's annual recurring revenue, but he described a company changing shape. "Enterprise was like 2% of revenue last year," he said. "This year, I would say like 15% of our revenue is enterprise contracts, and more than 50% of our revenue is B2B." Backstage, he said he expects more than half of revenue to come from enterprise and government customers in 2027.

That is the context for the whole episode. When your customers are hobbyists, the model behind the product is a matter of taste. When they are companies and government agencies signing contracts, the supplier you pick sets your margins.

Some offers on this page may be paid placements or contain affiliate links.