OpenAI Opens Its Agents API to Developers in Public Beta
OpenAI开放Agents API公测，AI智能体走向开发者
OpenAI has launched a public beta of its Agents API, letting developers build software agents that can plan and complete multi-step tasks on their own. The API, announced on September 10, provides tools for agents to browse the web, use apps, write code and handle files, with built-in guardrails and safety controls. OpenAI says the release aims to move AI assistants from answering questions to doing work. The company also introduced GPT-Live-1, a new voice model for real-time speech in the API. The moves come weeks after the launch of GPT-6 Astra, its most powerful model to date.
OpenAI于9月10日推出Agents API公开测试版，让开发者能够构建可自主规划并完成多步骤任务的AI智能体。该API提供了让智能体浏览网页、操作应用、编写代码和处理文件的能力，并内置安全护栏与控制机制。OpenAI表示，此举旨在让AI从回答问题走向真正执行工作。公司同时还发布了用于实时语音交互的新模型GPT-Live-1。这一系列动作发生在其迄今最强模型GPT-6 Astra发布数周之后。
