V4 flash is much better experience than glm 5.3 flash. But unusable for me as it's expensive

Long time api user. Recently it has started burning tokens a lott faster. I burn like a billion tokens a day so the cost change was noticable. Shifted to glm 5.3 flash as soon as I got to know about it - cheaper, smarter. Looks so good on paper. And maybe for some it is. But it is so slow. I get like ,42 tok/sec with glm and like 120tok/sec in v4 flash which is like night and day in user experience. Have to adjust. Pockets smaller than token requirement. What do you guys think.