OpenAI Agents API Enters Public Beta for Cloud-Based AI Agents

OpenAI has released the Agents API as a public beta for developers building cloud-based agents. According to the article, the API runs on the same infrastructure that powers Codex and ChatGPT. The core promise is practical: agents can run for hours, execute code, process files, and manage more complex workflows.

The API includes automatic context management, parallel tool use, and the ability to delegate tasks to sub-agents. It also supports MCP, custom functions, and built-in tools such as web search. For developers, that points to agent systems that can coordinate tools and subtasks without requiring every workflow step to be handled manually.

Developers can use OpenAI-hosted sandboxes or choose partner environments from Cloudflare, Vercel, and Oracle. The article states there are no extra fees beyond token usage, with billing based solely on tokens. That makes the initial cost model easier to evaluate for teams testing autonomous cloud agents.

The release gives developers public beta access to infrastructure associated with OpenAI’s existing agent-like products. Its focus on long-running execution, code, files, tools, and sub-agent delegation makes it relevant for teams exploring production workflows rather than simple chat interfaces. The practical takeaway: developers now have a clearer path to test agent systems on OpenAI’s platform while comparing hosted and partner sandbox options.