By skyforbes I am having trouble working with long mathematics proofs in Deep Think because it truncates the length of my input files to 200k tokens.
Deep Research is able to parse and manage or do it seems but I can’t tell if the reasoning is at the level of Deep Think.
I also work in Tex files or LaTeX and it doesn’t seem “supported” by docs by I tried pdf compiled version and Deep Think is truncating. It also truncated from beginning to end versus end to beginning and most of my major results are at end of long chains or proofs.
Is anyone else running into this issue? Or have a workaround that allows me to use the 10 daily Deep Think requests? I guess I could chop the bottom of the document but that is a lot of work for something OpenAI GPT Pro handles with ease. It used python to pull what it needs from my files as I suspect Gemini Deep Research might but not Deep Think.