Artificial intelligence · News
OpenAI Upgrades ChatGPT's GPT-6 Astra Pro With Parallel AI Agents
Published:
In brief
OpenAI announced on October 16, 2026 that GPT-6 Astra Pro inside ChatGPT now lets multiple AI agents work on separate parts of the same task at the same time, instead of one agent completing steps in sequence. The feature is currently limited to ChatGPT Pro subscribers, who can track each agent's progress and open finished work directly in the chat.
OpenAI said on October 16, 2026 that it had upgraded GPT-6 Astra Pro in ChatGPT, adding the ability for multiple agents to tackle different parts of a single task in parallel. The update also brings faster responses and improved use of plugins and external tools, according to the company's announcement, as reported by South Korea's AI Times.
How the parallel agents work
Previously, an agent inside ChatGPT would complete a complex request step by step. With the update, a task can be split into several independent subtasks that run at the same time, each handled by its own agent. OpenAI described the change in a statement quoted by AI Times: "multiple agents can tackle different parts of a task in parallel." Users can follow each agent's progress inside the conversation and open a completed task to review the results.
- Faster response times compared with the previous version of GPT-6 Astra Pro, per OpenAI's statement
- Improved use of plugins and external tools, according to AI Times and Lapaas Voice
- Multiple agents running on separate subtasks simultaneously instead of sequentially
- Users can monitor each agent's progress and open finished work within the same chat
AI Times gave a business-analysis report as an example: one agent could research a company's financials, another could analyze competitors, and a third could track market trends — all at once, rather than one after another.
Who can use it
According to AI Times and Digg, the parallel-agent feature is currently available to ChatGPT Pro subscribers in the Chat tab. OpenAI has not said whether or when it will extend to Plus, Business or Enterprise accounts, AI Times reported. Lapaas Voice, however, described the rollout as gradual, saying eligibility depends on a user's plan, account and workspace settings, among Pro, Business and Enterprise users specifically.
GPT-6 Astra, the underlying model family, is separately available within existing subscription allowances on ChatGPT Plus, Pro, Business and Enterprise plans, according to SessionWatcher. For Enterprise accounts, Lapaas Voice reported that access also depends on workspace administrator settings.
| Plan | Astra / Astra Pro access | Parallel agents |
|---|---|---|
| Plus | GPT-6 Astra included in allowance | Not reported as available |
| Pro | GPT-6 Astra Pro included in allowance | Available in Chat tab, per OpenAI/AI Times |
| Business | GPT-6 Astra Pro included in allowance | Gradual rollout reported by Lapaas Voice |
| Enterprise | Available, subject to workspace admin settings | Gradual rollout reported by Lapaas Voice |
Open questions and reaction
AI Times reported that OpenAI did not disclose how running multiple agents in parallel affects a subscriber's usage limits, nor did it specify a maximum number of agents that can run at once. Lapaas Voice noted that parallel agents do not guarantee faster results for every task: workflows that are inherently sequential, or that require heavy coordination between steps, may see little benefit.
Social-media reaction compiled by Digg was mixed. One comment cited by the outlet called the update an "amazing upgrade," saying it gets ChatGPT closer to "a powerful AI team working together," while Digg also reported that a large share of the reactions it reviewed were critical, with some users comparing GPT-6 Astra Pro unfavorably to competing models on speed.
Background
GPT-6 Astra is OpenAI's current model line inside ChatGPT, with a higher-tier "Astra Pro" version aimed at Pro, Business and Enterprise subscribers. SessionWatcher reported that OpenAI's API pricing for the Astra family stands at $10 per million input tokens and $50 per million output tokens, with a faster processing mode available at double that rate. These API rates apply outside of ChatGPT's subscription-based usage allowances.
Questions
- What is the GPT-6 Astra Pro parallel agents feature?
- It is an update to ChatGPT's GPT-6 Astra Pro model that lets several AI agents work on different parts of one task at the same time, rather than completing each step in sequence. Users can track each agent's progress and open finished work in the chat.
- Who can use the parallel agents feature right now?
- As of OpenAI's October 16, 2026 announcement, the feature is available to ChatGPT Pro subscribers in the Chat tab. Reports differ on whether it is also reaching Business and Enterprise users gradually, and OpenAI has not said if or when Plus will get it.
- Does this replace the GPT-6 Astra Pro model with a new one?
- No. Reports describe it as an upgrade to the existing GPT-6 Astra Pro model's agent capabilities, not the release of a new model.
- How much does GPT-6 Astra cost through OpenAI's API?
- According to SessionWatcher, API pricing is $10 per million input tokens and $50 per million output tokens, with a faster processing mode costing double that rate. This is separate from ChatGPT's subscription plans.
- Does running agents in parallel always finish tasks faster?
- Not necessarily. Lapaas Voice reported that tasks with clearly separable subtasks benefit most, while sequential workflows or tasks requiring close coordination between steps may see little speed improvement.
How the world is telling it
- KoreanSouth Korean technology outlet AI Times framed the update as a limited, Pro-only feature addition to an existing model rather than a new model launch, emphasizing that OpenAI had not disclosed how parallel agent use affects subscription usage limits or how many agents can run at once.
Sources
- AI Times — 오픈AI, 'GPT-6 아스트라 프로' 업그레이드...에이전트 병렬 실행 지원
- Digg — GPT-6 Pro in ChatGPT adds parallel agents for Pro users
- SessionWatcher — ChatGPT adds GPT-6 Astra to Plus, Pro, Business and Enterprise allowances
- Lapaas Voice — OpenAI Upgrades GPT-6 Astra Pro in ChatGPT With Faster Responses and Parallel Agents
This piece was prepared with artificial intelligence and passed an automated source check before publication. If you see an error, ask for a correction. Request a correction or reply · How we make this
Related
Artificial intelligenceArticle
How AI Algorithms Develop Bias, and What Safeguards Exist to Stop It
AI systems become biased when they learn from unrepresentative training data or encode human and institutional assumptions built into their design. Documented cases — an Amazon hiring tool and the 2018 "Gender Shades" facial-recognition study — show this can disadvantage women and people with darker skin. Regulators in the EU and standards bodies like NIST now require bias testing for high-risk systems.
Artificial intelligenceComparison
ChatGPT vs Claude vs Gemini in October 2026: models, prices and benchmarks compared
As of October 2026, ChatGPT runs on OpenAI's GPT-6 family (Astra for reasoning, Sol/Luna for everyday chat), Claude's flagship is Anthropic's Opus 5, and Google's generally available top model is Gemini 3.1 Pro while the newer Gemini 4 Argon is still in limited rollout. Entry-level paid plans cluster around $20 a month, but the three diverge sharply on context window, API cost and which benchmarks each company highlights.
Artificial intelligenceArticle
How AI Learns From Data: Training Explained Amid Meta's Scraping Clause Row
AI models learn by processing huge volumes of text, images or other data and adjusting internal parameters until their predictions match correct answers, repeated across billions of examples. Where that data comes from became a flashpoint in October 2026, when Meta's new commercial terms appeared to let it scrape business customers' websites for AI training, drawing criticism from publishers and legal scholars before Meta narrowed the wording.
Artificial intelligenceArticle
How AI Image Generation Works: Diffusion Models Explained
Tools such as DALL-E, Midjourney and Stable Diffusion create pictures using diffusion models: neural networks trained to turn an image into random noise and then reverse that process. To generate a new image, the network starts from pure noise and removes it step by step, guided by a text prompt, until a coherent picture appears.