Z.ai's GLM-5.3 Edges Mythos 5 on CyberGym
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Z.ai said its open-source GLM-5.3 model scored 84.5% on the CyberGym cybersecurity benchmark, edging Anthropic's restricted Mythos 5 which scored 83.8%.
- Z.ai will restrict GLM-5.3's most sensitive cybersecurity functions to verified users only, mirroring the gated-access approach used for Anthropic's Mythos 5.
Why it matters: Z.ai pairing a benchmark win with a verified-users-only restriction on sensitive capabilities shows a Chinese open-source lab adopting the gated-release pattern used for Anthropic's Mythos 5, narrowing the practical gap between open and closed frontier cybersecurity models while still keeping the riskiest features behind identity checks.
Ask SkimNews
