3 ms·They're mentioning using 20M tokens via z.ai subscription. GLM 4.7 is probably the model then.by PhilippGille 8mo agoThey're mentioning using 20M tokens via z.ai subscription. GLM 4.7 is probably the model then.