Multi-Agent (developers.openai.com)

OpenAI has introduced a new feature called Multi-Agent in its beta version for the GPT-6.1 Sol and GPT-5.6 models, which allows for the parallel execution of tasks by coordinating multiple subagents under a root agent. This capability is particularly useful for complex applications such as codebase exploration, documentation, and implementation, where breaking down tasks can lead to faster results and improved efficiency. The root agent can delegate specific tasks to various subagents that maintain their own context, minimizing interference and enhancing execution speed. The significance of Multi-Agent lies in its ability to streamline workflows by executing independent tasks simultaneously, offering significant performance improvements for research, analysis, and implementation projects. Developers can now utilize this feature to better manage workloads by enabling flexible task execution while being mindful of token usage and potential drawbacks for tasks requiring sequential operations. It provides a structured orchestration system that can simplify API integration with features for spawning agents and managing communication between them. This advancement in AI functionality promises to enhance collaborative efforts in tackling complex problem-solving tasks within the AI/ML community.