聲明:本文來自于（ID:ykqsd.com）授權轉載發(fā)布。

reported news highlights that openai has introduced new tools and api designed to streamline the creation of ai agents.

openai's key offerings this release include the responses api and agents sdk.

responses api combines the simplicity of chat completions api with the power of assistants api, offering developers a robust foundation for solving complex tasks through single api calls.

agentssdk is an open-source platform for multi-agent workflow management. it empowers developers to define agent roles and tools, enabling seamless task handoffs, while also providing safety checks and real-time monitoring to ensure efficient collaboration.

to enhance the practicality of ai agents, openai has provided three core tools: a web search tool, a file search tool, and a computer usage tool.

the web search tool supports models like gpt-4o and gpt-4o-mini, fetching real-time data from the internet and providing clear citation sources. this effectively equips agents with a dynamic knowledge base.

the file search tool utilizes vector storage and metadata filtering to quickly and accurately retrieve information from vast document libraries. for businesses, this is an invaluable resource for efficient knowledge retrieval.

the computer usage tool is equivalent to embedding an operator within the api, enabling complex operations like clicks, inputs, and drag-and-drop functionality. this breakthrough makes it possible for agents to interact with traditional systems in meaningful ways.

during a live broadcast, openai chief product officer kevin weil emphasized that 2025 will mark a significant milestone in the evolution of ai智能體 capabilities. he noted that chatgpt and the developer tools we've discussed will evolve from being mere question-answering tools to performing tasks that can genuinely impact the real world.