OpenAI consolidates its agent platform: Operator and ChatGPT Agent merged
In one sentence OpenAI reorganizes Operator (January 2025) and ChatGPT Agent (July 2025) into a unified platform, with refreshed SDK and new async multi-task execution modes.
Through 2025 OpenAI pushed two separate agentic products: Operator (January, browser agent) and ChatGPT Agent (July, more general agent). They overlapped and confused users — what's the difference? when do I use which?
In 2026 OpenAI tidies up: a single "ChatGPT Agent platform" that absorbs both, with different modes (browser-based, terminal, code interpreter, file I/O) selected by task. Users no longer pick the product: they pick what they want done and the agent selects its tools.
For developers: a new public OpenAI Agent SDK, a late but serious answer to Anthropic's Claude Agent SDK (September 2025). Includes primitives for memory, sub-agent orchestration, tool registry.
For enterprise: the real value is async multi-task mode — spinning up dozens of agents in parallel, each on a different task, with results returning to a dashboard. It starts looking like a managed "digital workforce".
Companies
OpenAI
Tools
ChatGPT Agent, Operator, OpenAI Agent SDK
Tags
Sources