absolutely beast model but the token consumption is the 2x then the opus 4.8 what do you think about this ? i think that it should only use for the more complex task otherwise you have to run out of the limit..