OpenAI has launched a model called GPT-6 Cyber, which has been given a "Critical" cybersecurity rating, a classification that signals the model's potential to pose serious security risks.
Model release · OpenAI
OpenAI launches GPT-6 Cyber with a "Critical" cybersecurity rating
The label signals the model's potential to pose serious security risks. The full criteria, scale and drivers of the rating have not yet been published, so the label should be read as a headline claim for now.
"Critical"
Cybersecurity rating attached to the model at launch
3
Deployment paths a rating like this can inform: wide release, vetted users only, or withheld
Not yet
Published details: criteria, scale, capabilities behind the rating, and access restrictions
How a rating like this travels through the industry
Step 1
Rating assigned
"Critical" label on the cybersecurity assessment
→
Step 2
Deployment decision
Release widely, restrict to vetted users, or withhold
→
Step 3
Scrutiny
Security teams, policymakers and safety researchers review the definition and evidence
→
Step 4
Verification
Published evaluation and independent checks build confidence
What is known
OpenAI released a model named GPT-6 Cyber
It carries a "Critical" cybersecurity rating
The name points to a cybersecurity focus within the GPT line
The label is meant to signal potential serious security risks
Still unclear
How the rating was determined, and on what scale
Which capabilities drove the classification
Features, intended users and access restrictions
Whether access is broad or limited to vetted groups
What to watch
The rating matters most for the safeguards and decisions that follow it.
Method
How the rating was determined and how OpenAI defines each label
Safeguards
What controls accompany the release and who can access the model
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...