Replit Shifts to OpenAI as Default Model Provider, Demoting Anthropic
Replit CEO Amjad Masad announced OpenAI's ChatGPT is now the default model on the platform, citing cost, performance, and partnership strength—though Anthropic models remain available as an option.
Last verified:
OpenAI Becomes Replit’s Default, Anthropic Demoted to Optional
Replit CEO Amjad Masad announced on October 17 at TechCrunch Disrupt that OpenAI is now the default large language model powering the cloud-based development platform, displacing Anthropic from that position. According to TechCrunch AI, while Anthropic’s Claude models remain accessible to users who explicitly select them, the shift to OpenAI as the default carries significant commercial implications—most developers do not manually override model selection and will default to the recommended option.
Masad framed the decision as driven by three factors: cost efficiency, measurable performance gains on both internal and public benchmarks, and partnership quality. “We publish a public benchmark as well, and consistently, Sol and Astra perform better on building vibe-coded applications from scratch,” Masad said, referring to OpenAI’s models. “It was a decision based on both cost, partnership, and performance.”
Replit’s Evolution Away from Anthropic
Replit’s shift underscores how model selection in developer tools has become competitive and economically material. According to TechCrunch AI, Replit was historically among Anthropic’s top-five customers and built its platform primarily on Claude before announcing an OpenAI partnership in August 2026. The company had continued promoting Anthropic models in production as recently as late September, with blog posts and social posts highlighting Claude Fable 5.1 and Claude Opus 5.5 capabilities.
TechCrunch AI noted that an unnamed source close to the Anthropic-Replit partnership disputed Masad’s characterization, claiming the partnership remained intact—a contradiction that suggests either ongoing discussions or messaging misalignment between the two organizations. Masad later clarified that Replit “remains model agnostic,” allowing users to select Claude or other supported models, but the default-selection mechanism creates a de facto hierarchy favoring OpenAI.
Enterprise Growth and Pricing Pressure
The timing of Replit’s shift aligns with the company’s reported growth in enterprise revenue. According to TechCrunch AI, enterprise contracts represented only 2% of Replit’s revenue in the prior year but grew to approximately 15% of revenue in the current period. This shift toward higher-margin enterprise deals may also inform the economics of model selection—enterprise customers often negotiate volume pricing and may benefit from vendor consolidation around a single primary provider.
The move signals that despite Anthropic’s strong technical reputation and earlier market position in developer tools, pricing and partnership dynamics remain determinative factors in model adoption within production platforms.
Why This Matters
For Anthropic, the loss of Replit as a top-five customer represents tangible revenue leakage in a segment where it previously held strong competitive position. For OpenAI, the default placement on a popular developer platform with substantial enterprise growth expands its distribution and embedding into workflows where model switching costs increase over time. For Replit users and enterprise customers, the shift creates a new baseline for code generation performance and cost—though the explicit option to use Anthropic models may preserve some negotiating leverage and prevents complete vendor lock-in.
The announcement also reflects a pattern in the AI software market: as model quality across leading providers converges on benchmark leaderboards, commercial terms, partnership depth, and default placement have become the decisive competitive axes.
Frequently Asked Questions
Can Replit users still access Anthropic's Claude models?
Yes. According to TechCrunch AI, Anthropic models remain available on the platform, but OpenAI's ChatGPT is now selected by default—a change that could substantially reduce Anthropic adoption since most users do not actively switch models.
Why did Replit switch to OpenAI?
Replit CEO Amjad Masad cited three factors: cost economics, superior performance on internal and public benchmarks for code generation tasks, and the strength of the partnership with OpenAI.
What was Replit's relationship with Anthropic before this shift?
According to the report, Replit was among Anthropic's top-five customers and built its platform primarily on Claude models before striking the OpenAI partnership in August 2026.