Replit Chooses OpenAI Over Anthropic for AI Infrastructure
AI coding platform Replit has decided to continue with OpenAI instead of Anthropic as its primary partner, based on cost and performance criteria.
AI coding platform Replit has announced that it is stopping its work with Anthropic as its primary partner and making OpenAI the default option.
Speaking at the TechCrunch Disrupt stage, Amjad Masad, CEO and co-founder of AI coding platform Replit, stated that the company is no longer working with Anthropic as its primary partner. Masad noted that this decision was made for various reasons including cost, partnership, and performance.
In the announcement, Masad stated that ChatGPT is now better, mentioning that they used an internal benchmark to evaluate metrics such as performance. Also stating that they published a public benchmark, Masad noted that the Sol and Astra models consistently perform better in building applications from scratch.
According to details clarified after the event, platform users will still be able to access Anthropic models, but OpenAI's ChatGPT model will now be offered as the default option. Replit had announced a partnership with OpenAI in August but had not given any indication that the partnership with Anthropic would be terminated. Before making the agreement with OpenAI, the platform relied on Anthropic's Claude models to run its infrastructure.