Comment by gpm

It's not a thinking model, and as such is probably going to be good and bad at different things. For the handful of prompts I've thrown at both, it (exp-1206) was generally better than gemin-2.0-flash-thinking-exp, but that could easily be a property of the prompts I tried.