
What happened
OpenAI released the Agents API as a public beta. It runs on the same infrastructure as Codex and ChatGPT, with automatic context management, parallel tool use, and task delegation to sub-agents.
Why it matters
Developers can now choose OpenAI-hosted sandboxes or partners like Cloudflare, Vercel, and Oracle, with no extra fees beyond token usage. This builds on the open-source Codex harness and supports MCP, custom functions, and web search.
What to watch
The test is whether developers adopt the partner sandboxes for production workloads, since billing is based solely on token usage. Watch how the public beta develops.
WHO IT HITSThird-party developers and enterprise engineering teams building cloud-based AI agents gain a new option for autonomous, long-running tasks with flexible sandbox hosting.
Ask the AI about this article →
Summaries like this, in your inbox every morning.
OpenAI is opening up the infrastructure behind its own flagship coding and chat products, Codex and ChatGPT, to outside developers. Until now, the same long-running agent capabilities that power those products were not directly available through a public API. Building on the open-source Codex harness means developers get a familiar foundation, while support for MCP, custom functions, and built-in tools like web search reduces the amount of plumbing they need to do themselves.
The choice between OpenAI-hosted sandboxes and partners like Cloudflare, Vercel, and Oracle gives teams flexibility in where their agents actually run. There are no extra fees beyond token usage, which simplifies cost planning for projects that might run agents for hours at a time. The public beta status suggests OpenAI is still gathering feedback before a wider rollout.
The stakes here hinge on whether third-party developers adopt the partner sandboxes for production workloads. If they do, it could accelerate the shift toward cloud-based agents that operate autonomously and delegate tasks to sub-agents. But because this is a beta with no additional fees, the real test will be reliability and scaling rather than pricing.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Dynatrace acquired Arize AI, adding AI observability, evaluation and agent monitoring to its application obser…
A Daily Dose of Data Science test kept LoRA adapters separate from a shared 7B base model, cutting 100 fine-tu…

A report by Spencer Kitts, Thomas Larsen and Sydney Von Arx says an OpenAI agent swarm very likely ran an atta…

Simon Willison wrote that many people, himself included, have gone through an existential crisis when a coding…

Stephen Aarons, a New Mexico defense lawyer of over 40 years, was held in direct contempt and fined $5,000 for…

Perplexity cofounder and Chief Strategy Officer Johnny Ho said GPT‑6 Astra can craft communications, edit real…
