GPT-6 Pro Gets Parallel Agents in ChatGPT: One Task, Multiple Agents, Progress Visible
OpenAI updated GPT-6 Pro in ChatGPT on October 16. The update is in the Chat tab and adds three things: faster responses, better use of plugins and tools, and parallel agents — the ability to split a single task across multiple agents working simultaneously.
GPT-6 Pro in ChatGPT is powered by GPT-6 Astra. It is distinct from GPT-6 Sol, which powers Instant through Extra High effort levels in the same model menu.
How parallel agents work
A user submits a task to GPT-6 Pro. The model identifies parts of the task that can run independently and assigns each to a separate agent. Users follow progress in the conversation view and can open each completed agent task to inspect its work.
OpenAI’s release notes do not specify a cap on the number of agents that can run simultaneously. No latency benchmarks were published. The October 16 announcement calls the improvement to plugin and tool use a separate line item from the parallel agent change, but gives no supporting data for either claim.
Who gets it
GPT-6 Pro is available on Pro $100, Pro $200, Business, and Enterprise plans. The October 16 update is rolling out gradually to eligible Pro users. In managed workspaces, access depends on workspace settings and rollout eligibility.
Usage limits apply even on the highest-tier plans. On Pro $200, hitting the weekly GPT-6 Pro limit downgrades the session to GPT-5.6 Thinking at Medium. On Pro $100, GPT-6 Pro and GPT-5.6 Sol Pro share a single weekly allowance.
Business Standard and Business Premium include 15 and 50 GPT-6 Pro messages per month and per week respectively, shared with GPT-5.6 Sol Pro. GPT-6 Pro is not available on Plus.
Ultra is different
OpenAI separately offers an Ultra mode in ChatGPT Work and Codex that also uses subagents for parallel work. Ultra and GPT-6 Pro in Chat are not the same feature. Ultra is a reasoning mode on a different ChatGPT surface; GPT-6 Pro parallel agents is an update to the standard Chat tab.
Context: agent competition intensifies
The parallel agent update arrived five days before Anthropic launched Claude Duet on October 21 — a proactive agent app built on Opus 5.5 that also runs background tasks autonomously. Both products reach roughly the same audience (high-end subscribers), take tasks off a user’s plate, and deliver results for review rather than autonomous action.
The structural difference is surface and pricing. GPT-6 Pro in ChatGPT’s Chat tab requires $100/month minimum. Claude Duet is included at $20/month Pro pricing. OpenAI has not announced a standalone Dots consumer agent app equivalent to Duet, though Dots was announced at DevDay for the Pro tier.
No benchmark data exists yet comparing the parallel agent implementation to alternatives. The absence of published quality or latency figures for the October 16 update is the key gap: OpenAI is shipping the architecture description, not the results.