Wire · technology
Experts Say There's Now an Open Source AI Model as Scary as Mythos
◆ Sectors
◆ Geography
◆ Source
◆ Verified
Fusion42 · 8 July 2026 · Fusion42 review
A Beijing-based AI company released GLM-5.2, an open-weight model matching Anthropic's restricted Mythos 5 in coding and vulnerability-detection capabilities, with no vendor guardrails—creating immediate cybersecurity risks as hackers rapidly adopt it on dark forums.
This Wire brief sits within Fusion42's coverage of AI Frontier Models and Cybersecurity.
◆ ◆ The Wire takeaway
Unrestricted open-source AI models now match frontier closed models in cybersecurity capability, collapsing vendor control and enabling adversarial use—founders in deeptech security, infra, and enterprise need to recalibrate threat models and competitive advantage assumptions immediately.
◆ Coverage
1 source · 8 Jul 2026
◆ Related on Wire
◆ Topics