OpenAI has started making a new model, GPT-6 Cyber, available for advanced security research through its Daybreak platform, initially to a select group of Daybreak Red customers.
OpenAI · Daybreak platform
OpenAI opens GPT-6 Cyber to a small group of security researchers
The security-focused model is available first to select Daybreak Red customers, with wider access planned over time. Further technical details are promised at a November 12 livestream.
Limited
Current access, restricted to select Daybreak Red customers
Nov 12
Date of the promised Cyber-related livestream and model update
Staged
Access expands gradually, after trusted users are observed
Staged-access rollout
Each step widens availability; only the first step is in place today.
Step 1 · Now
Select Daybreak Red customers
→
Step 2 · Planned
Expanded availability over time
→
Step 3 · Later
Broader release, once usage is observed
What is confirmed
The model is aimed at security research
Access is limited to a select group for now
Availability will expand over time
What is not yet known
Specific capabilities and performance benchmarks
Pricing
Criteria for selecting customers
What to watch on November 12
The livestream is expected to answer the open questions on the model.
Technical details
What the model can do
Availability terms
How access will expand
Safeguards
Measures to prevent misuse
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...