GPT-6 Pro in ChatGPT now splits tasks across parallel agents
Pro users can hand GPT-6 Pro a multi-part job in Chat, watch several agents work at once, and open each finished piece to check it.
If you are on ChatGPT Pro, pick GPT-6 at the Pro thinking level in the Chat tab and give it a task with separate parts. GPT-6 Pro can now send several agents at those parts at the same time, show each one's progress in the conversation, and let you open each completed task to see what it did.
What changed#
At 4:27 PM ET on Oct 16, @OpenAI posted that "GPT-6 Astra Pro in ChatGPT got an upgrade," with "faster responses and improved use of plugins and tools," and "multiple agents" that "can tackle different parts of a task in parallel." The ChatGPT release notes list the same three changes under "Updates to GPT-6 Pro model in ChatGPT."
Before: OpenAI's post presents parallel agents as new for GPT-6 Pro (powered by GPT-6 Astra) in Chat ("And now, multiple agents"). After: it can run several agents inside one Chat conversation. OpenAI gave no speed numbers and no limit on how many agents run at once.
Why it matters#
Our read: research with several independent threads (compare five vendors, check three sources, draft separate sections) no longer has to run one step after another in Chat. The review step matters as much as the speed: you can open each agent's finished task before trusting the combined answer.
How to use it#
These steps follow OpenAI's GPT-6 help article.
- In the model menu, choose GPT-6.
- Move the Thinking slider to Pro.
- Describe a task whose parts do not depend on each other, and say what each part should return.
- Follow each agent's progress in the conversation, then open completed tasks to check their work.
Our prompt (ours, untested):
Research these four project management tools in parallel: pricing, API limits, SSO support, and data export options for each. Work on each tool separately, then combine the results into one comparison table and note any field you could not confirm with a source.
Good to know#
- Rollout. The release notes say "eligible Pro users." Managed workspaces depend on "workspace settings and rollout eligibility."
- Limits apply. OpenAI's help article says: "A Pro subscription does not include unlimited use of the GPT-6 Pro model." On Pro $200, hitting the GPT-6 Pro weekly limit switches you to GPT-5.6 Thinking at Medium. On Pro $100, GPT-6 Pro and GPT-5.6 Sol Pro share one weekly allowance.
- No Intelligent UI at Pro. OpenAI's help article says the visual answers are "not available with Pro effort."
- Separate allowances. OpenAI's help article says "ChatGPT Work and Codex have separate allowances."
- Watch next: whether OpenAI publishes agent counts or speed figures, or brings this to Business and Enterprise.