Cyber Verification Program (expanded)
Anthropic launched an expanded Cyber Verification Program that folds in Project Glasswing, with three access tiers for vetted cyberdefenders using its latest Claude models.
The Cyber Verification Program and Project Glasswing are now one program with three tiers. Vetted security professionals get access to Anthropic's most capable models for cyber work with fewer safeguards. The tier details were not in the pages I could read.
- Date
- Tuesday, 6 October 2026
- Lab
- Anthropic
- Kind
- product
- Access
- closed API
Figures
| Measure | Value | Measured by |
|---|---|---|
| Verified vulnerabilities found by Anthropic, April to October | 5.5K+ as summarized by Techmeme from Reuters | company |
| Vulnerabilities found by Glasswing partners, April to July | 129K+ 33K+ were critical or high severity; as summarized by Techmeme from Reuters | company |
Access is by application and vetting. I could not read Anthropic's announcement, so the tier names and which models are covered are unconfirmed.
Sources
- www.techmeme.com/261006/p33#a261006p33
- www.techmeme.com/261006/p36#a261006p36
- www.theinformation.com/briefings/anthropic-expands-cyberdefenders-access-top-models
This record was partly confirmed: some claims could not be checked on 6 October 2026. How we check