One of the best-known AI coding platforms has changed its supplier, and not everyone agrees it happened. Replit chief executive Amjad Masad said onstage at TechCrunch Disrupt that his company no longer works with Anthropic, choosing OpenAI instead.
Masad gave three reasons: cost, partnership and performance. He argued that ChatGPT’s models, including ones he called Sol and Astra, consistently do better at building vibe-coded applications from scratch on both Replit’s internal tests and a public benchmark it publishes.
Replit had announced an OpenAI partnership in August without saying the older relationship was over. Anthropic’s Claude models had powered the platform during a stretch of rapid revenue growth, making the switch notable for startups watching model dependencies.
People familiar with the arrangement pushed back, telling TechCrunch the Anthropic partnership was still in place. Public evidence is mixed: a late-September blog post named Claude Fable 5 and 5.1 as models running in Replit Agent’s production environment, and a Replit employee posted about using Opus 5.5 around the same time.
Masad also shared growth details, saying enterprise moved from roughly 2% of revenue last year to about 15%, with more than half of revenue now coming from business customers.
The episode captures a new kind of churn risk for AI labs. Application startups can swap underlying models in weeks, so a partnership announcement is less a contract than a snapshot.
