Replit Shifts AI Partner From Anthropic to OpenAI

Replit, a popular AI coding platform for developers and startup founders, has reportedly ceased its collaboration with AI lab Anthropic. According to TechCrunch, Replit CEO and co-founder Amjad Masad announced the decision during TechCrunch Disrupt, explaining that the move was influenced by factors including cost and performance.

Masad stated that Replit's internal benchmarks, as well as public benchmarks, showed OpenAI's models, such as Sol and Astra, performing better for building "vibe-coded applications from scratch." Replit had previously used Anthropic's Claude models to power its cloud-based software development platform, a partnership that reportedly contributed to significant revenue growth for the company.

The CEO also mentioned that OpenAI was considered a "better partner for now." This shift comes after Replit announced a partnership with OpenAI in August, though the termination of the Anthropic collaboration was not previously indicated. However, TechCrunch noted some pushback from insiders, with one anonymous source suggesting the partnership with Anthropic was still active, and evidence from September indicating Replit was still using Anthropic products.

Masad also shared insights into Replit's business growth, noting a significant increase in enterprise revenue. Enterprise contracts now account for about 15% of the company's revenue, up from 2% last year, with over 50% of total revenue coming from B2B contracts. He also emphasized the importance of robust AI safety measures, describing the need for a "habitat" for AI agents that includes good orchestration, error recovery, security, and networking.

The AKUSO take: This move highlights the competitive landscape among AI model providers and the strategic decisions companies like Replit make based on performance and cost. It's a signal for developers and businesses about which AI tools are gaining traction in the real-world application space.