Replit Launches Proprietary Coding Model, Reduces OpenAI Dependency

Replit closed a $97.4 million funding round and released its own large language model for code, outperforming OpenAI Codex on key benchmarks. By building its inference layer in-house rather than relying on third-party APIs, Replit gains control over latency, model customization speed, and unit economics across millions of concurrent developer sessions.