SummaryCompaniesZ.ai says GLM-5.3 performed slightly better than Mythos 5 at finding software flawsIt still trailed Anthropic's model at developing working attacks from those flawsZ.ai plans public release after a two-week safety reviewFirm draws Mythos comparisons with initial access restricted to limited partnersBEIJING, Aug 14 (Reuters) - Chinese AI ​startup Z.ai said on Friday its open-source GLM-5.3 model had neared Anthropic's restricted Mythos 5 in identifying software vulnerabilities, bo...