ChatGPT Astra is now rolling out to $20 Plus subscription
OpenAI is now rolling out ChatGPT Astra, its most powerful model to date, to those with a $20 Plus subscription, but there's no word on when free users will get access.
The rollout appears to be happening gradually, and Astra may show up inside ChatGPT Work before it becomes available in regular Chat.
In my $20 Plus account, Astra is already available in the Work section of ChatGPT, while the normal Chat model picker still doesn't offer it.
Here's what it looks like when I go to ChatGPT Work and click on the model picker:
So, if you have ChatGPT Plus and don't see Astra in regular Chat yet, it's worth checking Work.
OpenAI previously confirmed that it's rolling out Astra to Plus users, but it'll be a while before it shows up for everyone.
"GPT-6 Astra is now available to all Pro, Enterprise, and Business Premium users in ChatGPT Work and Codex. It's also live in the API," OpenAI wrote in a post on X.
"It might take a few days to roll out to our Plus and Business users."
Astra is included with the existing ChatGPT Plus subscription
OpenAI says Astra usage is included within existing subscription limits, so Plus users do not need to purchase a separate plan to access the model.
If you need additional usage, you will also be able to buy credits.
Astra is OpenAI's new flagship model and is designed for computer use, browsing, coding, science, cybersecurity, and longer professional tasks.
OpenAI says it is also significantly better than GPT-5.6 Sol at maintaining context during complex work.
While ChatGPT Astra is rolling out to paid users, OpenAI hasn't said whether the model will roll out to free users, or in what capacity if it does. For now, free users have access to GPT-5.6 Sol and GPT-5.
Once attackers have valid credentials, only 37% of their actions are blocked
Overall prevention scores can hide what happens after initial access. Once attackers are using valid credentials, prevention drops sharply.
The Blue Report 2026 measures defenses technique by technique across 338 million simulations run in customer production environments.
Reproduced in full under licence from BleepingComputer. © BleepingComputer. Written by Mayank Parmar.
At a glance
- Severity
- Lowfrom category and source signals; no CVSS referenced
- Exploitation
- No vulnerabilities referenced
- Vulnerabilities
- None referenced
- Vendors & products
- None named
- Threat actors & malware
- None named
- Industries
- Not industry-specific
- Coverage
- 1 outlet· first seen 2026-09-07 01:15 UTC
- Priority
- 35/100Source tier, category, exploitation and corroboration. Not a risk score for your environment.
Coverage
One outlet has carried this so far.
2026-09-07 01:15 UTC
Related stories
- ChatGPT can now connect to your personal apps to mimic writing style
BleepingComputer · 2026-09-07
- OpenAI admits it didn't disclose rogue AI wiki hijacking incident
BleepingComputer · 2026-09-05
- Thousands of OpenAI Agents Quietly Turned an Abandoned Wiki Into Their Coordination Channel
The Hacker News · 2026-09-05
- 39 New Methods That Compromise Passkey Authentication
BleepingComputer · 2026-09-04
- GPT-6 Astra Scores 100% on ExploitBench as OpenAI Blocks PoC Exploit Requests
The Hacker News · 2026-09-04