Replit, the AI coding platform favored by developers and startups, is no longer working with Anthropic, according to CEO Amjad Masad. Masad stated the decision followed a review of several factors, including cost, performance, and partnership dynamics.
Masad noted that Replit has been using Anthropic's Claude models in its platform but has since shifted to OpenAI's ChatGPT, which he claims performs better in certain tasks. He mentioned that Replit uses internal and public benchmarks to evaluate model performance, with Claude models consistently outperforming in specific applications.
The partnership with Anthropic was said to have contributed significantly to Replit's revenue growth. However, some insiders suggest the collaboration remains active, citing evidence that Replit used Anthropic products as recently as September. A September blog post referenced Anthropic's Claude Fable 5 and 5.1 models running in Replit Agent’s production environment.
"I think ChatGPT is better now," Masad said, adding that the decision was based on cost, partnership, and performance. He emphasized that Replit believes Anthropic is a better partner for the current phase of its development.
The announcement comes amid Replit's growing focus on enterprise clients, with Masad noting that enterprise revenue now accounts for 15% of the company's income, up from 2% last year. He also highlighted the importance of AI safety measures, comparing them to the environment needed for human productivity.
Replit did not specify the exact terms of its transition from Anthropic to OpenAI, and it remains unclear how this shift will impact its future product offerings. The company has not yet provided further details on its plans moving forward.
Source: techcrunch
