Inspect the documented capabilities, access conditions and sources for your task.
Initial research: 2026-09-21. Check current terms before choosing.
Updated
The facts
- Version
- public beta
- Price
- Price not verified. Open the tool record for the available pricing sources.
- API
✓Supported - Open weights
✕Not supported - Commercial use
- yes
- Computer use
✓Supported Exposes the same harness that runs Codex and ChatGPT for Work; OpenAI describes GPT-6 Astra as state of the art at computer use
Source for Computer use (opens in a new tab)- Browser control
✓Supported ChatGPT Agent, which absorbed Operator in July 2025, runs its own virtual computer with a browser
Source for Browser control (opens in a new tab)- Multi-step autonomy
✓Supported Long-lived sessions running for hours or days, with subagent coordination and context compaction across long tasks
Source for Multi-step autonomy (opens in a new tab)- Sandboxed execution
✓Supported OpenAI-hosted sandbox, private VPC, or 9 partner sandboxes including Vercel, Cloudflare and E2B
Source for Sandboxed execution (opens in a new tab)
Worth knowing
OpenAI shuts down Agent Builder, the Evals platform and the reusable prompts API on 30 November 2026, announced 3 June 2026.
From the publisher
Automatic source checks. These do not change the verification date of the facts above.
Sources
Beyond feature support
Quality evidence.
How well a model performs depends on the task, version, and settings.
We have not published a reviewed quality evaluation for these versions yet. Feature support and price do not establish output quality.
Try the same small task with each option. Check accuracy, consistency, and how much editing you need to get a usable result.