I don’t have api access to the GPT o1 preview model… yet. But acting as a messenger between the comic generator and ChatGPT I was able to generate a comic with the new model. While the result wasn’t any more creative than anything GPT4 would do, it did output a usable comic on the first try.
One other noteworthy thing is an apparent continuity in the background images. In the first two background descriptions, the model made a point to repeat that there are jack-o-lanterns on the sidewalk. This was an unnecessary part of the scene, but would be appropriate in a sequential story.
