News · Explainer · 3 min read
What "parallel agents" means, now that ChatGPT's GPT-6 Pro has them
OpenAI says GPT-6 Pro in ChatGPT can now split a task among several agents working at the same time. Here is what that means in plain terms.
On October 16, 2026, OpenAI’s ChatGPT release notes said that GPT-6 Pro in the chat tab now offers “multiple agents that can work on different parts of a task in parallel.” Users can follow the agents’ progress in the conversation and open completed tasks to see what each agent did.
The same update listed faster responses and improved use of plugins and tools. OpenAI said it is rolling out to eligible Pro users, and that availability in managed workspaces depends on workspace settings.
OpenAI’s help center describes the Pro option as meant for “difficult tasks and longer-running workflows,” and says GPT-6 Pro is powered by a model called GPT-6 Astra.
What an agent is
An AI agent is a model that works toward a goal in steps. It makes a plan, uses tools such as web search, looks at the results and decides what to do next.
What “in parallel” adds
In parallel means at the same time. A single agent handles the steps of a job one after another. With several agents, a job that splits into independent parts can be worked on all at once. As a general example, a request to compare five products could give one product to each agent, with the findings combined at the end. OpenAI’s notes do not describe how GPT-6 Pro divides its work.
Why it matters
Splitting a job can shorten the wait when its parts do not depend on each other. In general, it also means more model work happens behind one answer, because each agent reads and writes its own text. The option to open each finished task gives you a way to check what each agent did. The Agents at Work course explains how agents plan, use tools and hand work back to you.
How parallel agents can split a job
- 1One requestFor example, compare five products.
- 2Split into independent partsThe job only splits well when its parts do not depend on each other.
- 3Several agents work at the same timeOne product to each agent. Each plans, uses tools and reads its own results.
- 4Findings combinedThe results come together in one answer, and you can open each finished task to check it.
Sources
Go deeper · guided courseAgents at WorkWhat an AI agent is, where it helps, and how to brief one for a real task.Starts with: From chat to agent · 11 minStart the course →