Hey my friends, 

There is a new version of the open source video AI model Wan. And so i digged into it, and compared the three main methods for image to video. This time there is no new workflow included. These workflows are all publically available. And there is little that i could add to it.

More important and of interest is the comparison between the 14B model with two ksampler, the 5B Model with one ksampler. And the GGUF version.

The youtube explain video can be found here: https://www.youtube.com/watch?v=mlK7jFhBd60