yes but as far as i know gpt tokenizer is about the same as opus 4.6's, where 4.7 is seeing something in the ballpark of a 30% increase. this should still be cheaper even disregarding the concerns around 4.7 thinking burning tokens