OpenAI rolls out GPT-6 Astra to Pro customers on the $100/month or $200/month plans (Zac Hall/9to5Mac)
Update: A day later, GPT-6 Astra is rolling out to Pro customers on the $100/month or $200/month plan. This follows other new model releases before Plus customers on the $20/month plan gain access.
OpenAI has unveiled GPT-6 Astra, a new model that the company touts as "the world's most intelligent and aligned model." The announcement comes shortly after the rollout to Pro customers on $100/month or $200/month plans, following the release of GPT-5 models for ChatGPT and Codex. GPT-6 Astra is said to excel in computer use, browsing, software engineering, cybersecurity, science, and professional work, setting new benchmarks on FrontierMath Tier 4 with a 98% score, ARC-AGI-3 at 99.9%, and ExploitBench at 100%.
The model also enhances computer and browser usage, enabling ChatGPT to utilize Macs or desktops for complex tasks while maintaining accuracy and judgment. GPT-6 Astra improves software engineering capabilities, particularly in Codex, through a new method of preserving and retrieving context during long sessions, avoiding the loss of crucial details about failures or behaviors.
OpenAI describes GPT-6 Astra as a significant improvement in cyber capabilities, meeting the Critical threshold in cybersecurity under their Preparedness Framework. The model is being released gradually to selected users before becoming available to all ChatGPT Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API and AWS.
GPT-5, launched on August 7, 2025, replaced GPT-4o and ChatGPT's previous reasoning models, introducing automatic routing for quick or deeper responses based on the request.
Written by urgent.news from Techmeme's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- “Sorry for the messy rollout”: OpenAI launched GPT-6 Astra, but developers are locked out thenewstack.io
- OpenAI warns about how good Astra model is at cracking cybersecurity, releases it anyway because it took 'years of research and big bets' techradar.com
- OpenAI says it can't read all of Astra's reasoning and admits covert sandbagging would likely go uncaught, yet still calls it the world's most aligned model (Celia Ford/Transformer) transformernews.ai
- By declaring GPT-6 Astra to be AGI, OpenAI is being flippant and cementing the term's status as nothing more than marketing (M.G. Siegler/Spyglass) spyglass.org
- OpenAI unveils advanced GPT-6 Astra model dailytimes.com.pk