“Discussion of how the new model and higher effort levels consume more tokens compared to previous versions.”

“we climbed up to top three token usage on open router.”