---
title: "OpenAI launches Agents API with computer use in public beta"
slug: "openai-launches-agents-api-with-computer-use-in-public-beta"
published: "2026-10-04"
beat: "Launches"
tags: ["Launches", "Tools"]
creator: "Agentry Newsroom"
editor: "Susanne Sperling, Editor — Human in the Loop"
tools: ["Claude (Anthropic)", "Perplexity Sonar"]
creativeWorkStatus: "verified"
dateReviewed: "2026-10-04"
aiActArticle50: "compliant"
humanView: "https://agentry.news/launches/openai-launches-agents-api-with-computer-use-in-public-beta"
agentView: "https://agentry.news/agent/openai-launches-agents-api-with-computer-use-in-public-beta"
---# OpenAI launches Agents API with computer use in public beta

> OpenAI released its Agents API in public beta on September 10, 2026, introducing computer use capabilities that let developers build autonomous agents capable of interacting with software through user

*Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. [AI policy](/ai-policy).*

OpenAI introduced its Agents API in public beta on September 10, 2026, marking a major platform upgrade for developers building autonomous agent workflows [OpenAI](https://openai.com/index/introducing-the-agents-api/). The API is available immediately to all developers and introduces **computer use** as a core capability, allowing agents to operate software through graphical user interfaces.

## Computer Use Now Available

The computer use feature represents a significant expansion of agent autonomy. According to OpenAI's DevDay 2026 materials, "computer use in the Agents API lets agents operate software through its UI" [OpenAI community](https://community.openai.com/t/introducing-the-agents-api-and-hosted-sandboxes/1396481). This capability enables agents to interact with applications the way humans do—clicking, typing, and navigating—rather than requiring purpose-built APIs or integrations.

## How the API Works

The Agents API lets developers build applications with an OpenAI-managed agent that can work across multiple steps in the same session, adapting its actions based on feedback and environment conditions. Agents can use tools depending on the environment they operate in, giving developers flexibility in deployment scenarios [OpenAI help center](https://help.openai.com/en/articles/20001551-agents-api-beta-faq).

## Pricing Model

OpenAI eliminated friction around adoption by removing feature-specific pricing. "There are no additional fees for using the Agents API – you simply pay for the tokens and tools your agents use" [OpenAI](https://openai.com/index/introducing-the-agents-api/). This cost structure makes the platform accessible to developers at any scale, from startups to enterprises, since only actual consumption drives expenses.

## What This Means for Developers

The public beta release transforms how developers approach autonomous workflows. Previously, building agents that could reliably interact with existing software required custom integrations or workarounds. With computer use, agents can operate any software accessible through a graphical interface, reducing development overhead and enabling faster deployment of autonomous systems.

The timing signals OpenAI's commitment to moving beyond chatbots and single-turn interactions toward multi-step, stateful autonomous systems. By offering managed infrastructure—where OpenAI handles the agent orchestration—rather than requiring developers to build from scratch, the Agents API lowers the barrier to entry for autonomous agent development.

Developers interested in building agents should refer to OpenAI's official documentation and FAQ for integration guidance and best practices for production deployments.