Shared Article
OpenAI has launched a WebSocket mode for its Responses API to support long-running, tool-heavy agentic workflows with lower latency. This persistent connection reduces overhead and can speed up end-to-end execution by up to 40% for complex tasks.