Woke up and saw Gemini pushed out another new video model, text-to-video jumped straight to #1. I always just glance at leaderboards and move on; when you’re actually doing the work, the gap between #1 and #3 probably matters less than whether your prompt is any good.
Still, for people editing footage, top vendors leapfrogging each other is a good thing — this time last year we were still worrying about faces melting in a 3-second shot, now the conversation is about camera moves and consistency.
What I actually want to know is the max duration, whether it can hold the same character across a clip, and how long one generation takes. Guess I’ll wait until I can actually try it, the official demos are always cherry-picked.
