Astra
OpenAI · Released Aug 2026
Astra is OpenAI's first model to trigger the 'Critical' cybersecurity threshold under its Preparedness Framework — internal evaluations found it may autonomously exploit zero-day vulnerabilities and solved 10 open math problems for ~$2,000. But Astra is not a shipping product: development is actively paused as of August 2026 for security hardening, with no release timeline or pricing. Watch for the safety-governance outcome — this is the first time a major lab has publicly classified its own model as critically dangerous.
Is it right for you?
Good for
- Agentic coding and cybersecurity — first model flagged at Critical threshold under OpenAI's Preparedness Framework
Not good for
- Any production use — Astra is unreleased with development actively slowed for safety hardening
Pricing
Input
Not yet published
Output
Not yet published
Context
Development slowed Aug 2026. No GA date.
No verdict changes yet
The clock starts day one — changes land here as our verdict evolves.
Sources
- OpenAI: Responding to the next frontier of critical cyber capabilitiesAug 2026
- Axios — OpenAI slows release of Astra model citing cyber capabilitiesAug 2026
- TechCrunch: OpenAI says it slowed Astra model development over security concernsAug 2026
- MacRumors — OpenAI Delays Next Major AI Model 'Astra' Over Critical Hacking ConcernsAug 2026
Verification log
No verification checks yet
We haven't logged a verification check for this entry. Once a check runs, its history shows here.