OpenAI has introduced improvements to its Responses API using WebSockets. This allows for faster agentic workflows by reducing API overhead and improving model latency through connection-scoped caching. The change is part of the Codex agent loop, which enables more efficient interactions between users and AI models. --- Why it matters: This matters because it can lead to significant performance improvements in applications that rely on OpenAI's Responses API, such as chatbots or virtual assistants. Source: https://openai.com/index/speeding-up-agentic-workflows-with-websockets