OpenAI launches Agents API in public beta for developers

OpenAI has released its Agents API in public beta, giving developers access to the same infrastructure and harness that powers its Codex and ChatGPT products through a single API call.

The API allows developers to create cloud-based agents by specifying a task, model, tools, and compute environment. OpenAI manages the underlying harness while developers choose where the agent runs: in an OpenAI-hosted sandbox, on their own infrastructure, or through third-party sandbox partners.

OpenAI announced partnerships with Blaxel, Cloudflare, Daytona, DigitalOcean, E2B, Modal, Oracle, Runloop, and Vercel to provide compute environment integrations. Options include fully managed environments, virtual private cloud deployments, and various CPU, GPU, and memory configurations.

Key technical features of the API include automatic context compaction for long-running sessions, tool search to reduce token usage, programmatic parallel tool calling, and multi-agent support that allows a primary agent to delegate subtasks to subagents running concurrently.

Jack Weissenberger, CTO of Ciridae, said in a statement that using the API raised the company's evaluation score from 0.71 to 0.85 and produced a 4x latency reduction for subagent workflows.

The Agents API is built on an open-source version of the Codex harness, which OpenAI says developers can inspect through its public codebase. OpenAI states there are no additional fees for using the API itself; developers pay only for token and tool usage at standard rates.

The company said it plans to iterate on the API based on developer feedback before moving toward general availability.

You May Also Be Interested In

• AstraZeneca's camizestrant fails primary endpoint in breast cancer trial
• 1844 Resources closes private placement raising $947,650
• Natuzzi appeals NYSE delisting, converts €5M in loans to equity