On ChatGPT Pro? GPT-6 Pro can now split one task across parallel agents, so give it work with separate parts

#What changed

OpenAI says GPT-6 Pro in ChatGPT's Chat tab got an upgrade on October 16. It now answers faster and does a better job using plugins and tools to carry out a task. The bigger change is that it can now run several agents on different parts of one task at the same time. You can follow each agent's progress in the conversation and open a finished task to see what that agent did. OpenAI first posted it as "GPT-6 Pro" and then edited the post to "GPT-6 Astra Pro". The ChatGPT release notes call it GPT-6 Pro, powered by GPT-6 Astra.

#Who is affected

ChatGPT Pro subscribers who pick the Pro thinking level in Chat. GPT-6 Pro is also listed for Business and Enterprise plans, but this upgrade is rolling out to eligible Pro users first, and availability in managed workspaces depends on workspace settings. Plus doesn't include the Pro level in Chat. This change is to Chat. ChatGPT Work and Codex have their own models and allowances.

#What to do now

• To use it, pick GPT-6 in the model menu and slide Thinking to Pro. If you don't see it on desktop, use Menu > Check for Updates and install the latest version.
• Give it work with parts that can run independently, like comparing several vendors, checking a few data sources, or drafting separate sections. Name the parts and say what each one should return, so the agents don't overlap.
• Ask for a final combined answer that cites which part each finding came from. Then open the completed tasks and spot-check them before you rely on the summary.
• Watch the limit. GPT-6 Pro isn't unlimited on any Pro plan. On Pro $200, reaching the 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, so save Pro for tasks that actually need it.
• Don't expect Intelligent UI layouts here. OpenAI says the Pro level doesn't support them, so you'll get mostly text.

#What is not confirmed

OpenAI hasn't said how many agents can run at once, whether a parallel run uses more of your weekly Pro allowance than a normal Pro answer, or when managed workspaces get it. The advice to split work into named parts is our suggestion. We haven't tested whether it improves results. The rollout is gradual, so you may not have it yet.

#Sources

• OpenAI on X (edited post): https://x.com/OpenAI/status/2111192259073094035
• OpenAI on X (original post): https://x.com/OpenAI/status/2111185504746213705
• ChatGPT release notes, "Updates to GPT-6 Pro model in ChatGPT": https://help.openai.com/en/articles/6825453-chatgpt-release-notes
• GPT-6 and other models in ChatGPT (picker, Pro limits, Intelligent UI): https://help.openai.com/en/articles/20001354-gpt-6-and-other-models-in-chatgpt