Does anyone else feel like with each new model and change, the performance and output quality just keep getting worse? I personally think Google AI Studio was at its best, most productive, and highest quality between February/March and August 2024 with the 0827 Experimental model. Anyone else agree?
4 Likes
yeah totally. i feel like the 0827, 1216, and 1106 exp models were some of the best, at least for my usage cases, especially 08 and 11. the newer models obviously are better for some other cases but i dont typically use them as much for those.
Yes and No, the models take a while to stabilize, the newer models appear to be catching up quicker than the old models. Deep Research and 2.0 Flash seem to be improving quicker than the others, probably due to the higher usage. This is only based on the user-side observations, I am only beginning. Hope you don’t mind my chiming in.