Z.ai says viral 'Ox Alpha' was GLM-5.3-Flash, served on Chinese chips

Z.ai confirmed the anonymous "Ox Alpha" model that topped OpenRouter's usage charts is GLM-5.3-Flash, a 320B-parameter mixture-of-experts system with 18B active parameters and a one-million-token context. The lab published the weights under an MIT license and said the stealth run, roughly 62 trillion tokens, was served entirely on domestically produced Chinese chips. Zhipu's Hong Kong shares rose more than 12%.