Chinese startup Z.ai to publicly release GLM-5.3, challenges Mythos 5
Chinese startup Z.ai said it will release GLM-5.3 publicly in about two weeks, aiming to compete with the more locked-down Mythos 5 from Anthropic, especially in cybersecurity.
In Z.ai's own tests, GLM-5.3 edged out Mythos 5 at spotting software bugs (scoring 84.5% vs. 83.8%), but fell behind when it came to converting those bugs into working attacks.
Z.ai trusted access raises expert concerns
Z.ai plans to release GLM-5.3 publicly soon, while its most sensitive cybersecurity functions would be available only to verified users through a "trusted access" program, with built-in safeguards meant to prevent misuse, while still letting legitimate users do things like debugging and security testing.
Still, some experts worry that once it's out there, keeping the tech from being used for illicit purposes could get tricky, even with all the protections in place.