# Replit Ends Partnership with Anthropic in Favor of OpenAI CEO Amjad Masad cites performance and cost as primary drivers for the strategic shift. - **Source:** [TechCrunch](https://techcrunch.com) - **Published:** 2026-10-16 - **Topic:** [Business](https://brifing.si/c/business) - **Original article:** - **On Brifing:** https://brifing.si/s/1fb82cfd-8ff4-4d8d-b480-9dfdd8bb49cf ## Strategic Partnership Pivot Replit CEO Amjad Masad announced at TechCrunch Disrupt that the company has officially ceased its partnership with Anthropic. This decision marks a significant shift for the AI-powered coding platform, which previously relied on Claude models to support its development environment. The company is now transitioning its infrastructure to prioritize OpenAI's technology stack. ## Performance and Cost Metrics Masad justified the move by citing internal benchmarks that favor ChatGPT over Anthropic's offerings for specific coding tasks. He noted that the decision was driven by a combination of cost efficiency, partnership alignment, and overall model performance. Replit claims its internal metrics show superior results for vibe-coded applications when using OpenAI models. ## Contradictory Internal Reports Despite the CEO's public statement, sources familiar with the situation suggest the partnership may still be active. Evidence from a September blog post indicates that Replit Agent was utilizing Claude Fable 5 and 5.1 models in its production environment. This discrepancy between executive messaging and operational reality remains a point of confusion. ## Enterprise Revenue Growth During the event, Masad highlighted a major shift in the company's revenue composition. Enterprise contracts have grown from 2% of total revenue last year to 15% in the current year. Furthermore, B2B services now account for more than 50% of Replit's total revenue, signaling a successful pivot toward professional and corporate clients. ## Defining AI Habitat Requirements Masad emphasized the necessity of creating a secure habitat for AI agents to function effectively. He compared this to human office requirements, stating that agents need proper orchestration, error recovery, and sandboxing. These infrastructure components are essential for deploying AI safely within a production environment, ensuring identity and authorization are strictly maintained. ## Historical Context of Collaboration The previous collaboration with Anthropic was widely credited with helping Replit scale its revenue significantly during its growth phase. By integrating Claude models directly into the browser-based development platform, Replit provided users with advanced coding assistance. This transition away from Anthropic represents a major change in the platform's underlying technical architecture. ## Operational Transparency Challenges The ambiguity regarding the current status of the Anthropic partnership highlights the complexities of managing AI-dependent software platforms. While the CEO publicly declared the end of the relationship, the continued presence of Anthropic models in recent production environments suggests a phased migration or potential internal misalignment regarding the timeline of the transition. ## Future Strategic Outlook As Replit continues to scale its enterprise offerings, the choice of AI model provider will remain a critical factor in maintaining performance standards. The company's focus on B2B revenue suggests that future infrastructure decisions will be heavily influenced by reliability, security, and the ability to support complex, production-grade applications for corporate users. ## Sources covering this story - https://techcrunch.com/2026/10/16/replit-is-no-longer-working-with-anthropic-ceo-says/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._