Since 3 Pro we already know that Google's model are very good in benchmarks, but when you start programming with them... not really good and each model has its quirks that make me continue the project with something else, once I fixed gemini's bugs.
Since 3 Pro we already know that Google's model are very good in benchmarks, but when you start programming with them... not really good and each model has its quirks that make me continue the project with something else, once I fixed gemini's bugs.
I was having immense fun wit 2.5 Pro. Going as far as copying the whole repo into context with repomix.
I don't know what happened at some point, they started benchmaxxing and using the model for real programming just became impossible, probably 3.1 Pro is still "hitting a snag oh boy" in loop since the last time I tried it for a new project from scratch.