← Back to live feed1 story
Cybersecurity organizations and independent researchers can now apply for early access to test the newest model from Z.ai for potential vulnerabilities. The company is seeking collaboration to conduct security testing and evaluation intended to harden the system through external review.
The initiative follows the release of GLM-5.3, which posted a score of 84.5 on CyberGym to lead in vulnerability discovery. Z.ai is requiring interested parties to complete an application form to obtain early access for the testing phase.
Sign in to suggest edits
Key sources
- SOURCE@zixuanli_“If you’re interested in early access for security testing and evaluation, please complete the form below”x.com
- SUPPORT@zai_org“invite cybersecurity organizations and security researchers to collaborate with us on evaluating GLM-5.3”x.com
- SOURCE@zixuanli_“OpenVuln will scan the repository for potential vulnerabilities and help you safeguard your project”x.com
- SUPPORT@zixuanli_“It found a potentially serious vulnerability in Cursor”x.com
- SUPPORT@arena“post-training on the 743B base model”x.com
- SUPPORT@mtslive“found 2,436 vulnerabilities across 269 projects, 1,097 critical or high”x.com
- SUPPORT@semianalysis_“massively beats every American open model: Nemotron, Laguna, Inkling”x.com
- SUPPORT@genaispotlight“deployed directly on Hugging Face Spaces and GitHub under Apache 2.0”x.com