[anabranch]

I wanted to better understand the potential impact for the tokenizer change from 4.6 and 4.7.

I'm surprised that it's 45%. Might go down (?) with longer context answers but still surprising. It can be more than 2x for small prompts.

[pawelduda]

Not very encouraging for longer use, especially that the longer the conversation, the higher the chance the agent will go off the rails