Replit makes OpenAI its default over Anthropic

Replit has made OpenAI its default model provider, replacing Anthropic, according to CEO Amjad Masad, who discussed the change over the weekend. The shift does not remove Anthropic’s Claude from the platform: customers can still access Claude. Instead, the change concerns which model is used by default when users do not choose another provider. That distinction is important because defaults often shape what most developers actually use. For AI coding platforms, default routing is a powerful lever. Many users accept the suggested model, especially when they are focused on building software rather than comparing benchmarks. A default can therefore influence adoption, feedback loops, and revenue across the model ecosystem. Replit’s decision gives OpenAI a prominent position in a popular coding environment and signals that competition among AI labs is increasingly about distribution as well as raw model quality. The switch also reflects shifting enterprise relationships among AI labs and coding-tool vendors. Model providers are competing for integrations that put their systems in front of developers at the moment they write, test, and debug code. Reliability, latency, cost, and features such as tool use and long-context coding all matter. A provider that becomes the default may gain more usage data and more opportunities to improve its products, while rivals must rely on users actively selecting them. For Replit customers, the immediate effect may be subtle but practical: new projects and prompts could route to OpenAI models unless they opt for Claude or another option. The company’s choice may also encourage other platforms to re-examine their own default arrangements. The larger lesson is that in the AI coding market, being available is not the same as being the default. Defaults can determine which models become habitual, and habits can be hard to reverse. Replit’s move is therefore a competitive signal, not simply a product setting.