From TechCrunch · Kirsten Korosec, Lucas Ropek, Rebecca Bellan · · 2m
Replit drops Anthropic for OpenAI as default
The AI coding platform switches its primary model partner, citing performance, cost and partnership strength.
In brief
The AI coding platform switches its primary model partner, citing performance, cost and partnership strength. Replit has made OpenAI its default AI model over Anthropic, though users can still access Claude, potentially cutting spending with its former top-five customer relationship. Originally reported by TechCrunch.
Replit switches to OpenAI
Replit, the AI coding platform used by startup founders and developers, is no longer working with Anthropic as its default model partner. CEO Amjad Masad announced the shift Thursday at TechCrunch Disrupt.
Three reasons for the change
Masad cited cost, partnership strength, and performance. He said ChatGPT is now better, noting that Replit uses internal benchmarks where Sol and Astra perform better on building vibe-coded applications from scratch.
“I think ChatGPT is better now”
Anthropic not completely cut out
Customers can still access Anthropic models on Replit's platform, but OpenAI's ChatGPT is now the default. Most users stick with defaults rather than selecting models themselves, likely reducing Replit's Anthropic spending.
Replit was top Anthropic customer
Before partnering with OpenAI in August, Replit relied on Anthropic's Claude models. Masad noted that Replit had been a top-five customer of Anthropic's until recently, and that partnership helped the company grow revenue significantly.
Masad clarifies model choice
After Disrupt, Masad told TechCrunch that Replit remains model agnostic. Users can choose Anthropic's Claude or other supported models, but OpenAI is now the preferred partner for powering Replit Agent based on performance, economics and…
“Replit remains model agnostic”
Insiders dispute the timeline
An unnamed person familiar with the partnership said it was still in place. Evidence on Replit's website and social media from late September shows the company promoting Anthropic's Claude Fable five and five point one…
Enterprise revenue growing fast
Masad shared that enterprise contracts grew from two percent of revenue last year to fifteen percent this year.
AI agents need safe habitats
Masad discussed AI safety, comparing agents to humans needing proper environments. He said agents need good orchestration, error recovery, safety, security, sandboxing, networking and identity authorization to deploy safely and do good work.
“agents need a habitat for them to exist and do good work”
What matters most
Replit has made OpenAI its default AI model over Anthropic, though users can still access Claude, potentially cutting spending with its former top-five customer relationship.
