The models behind an Autotube video.
One page per model, with the capabilities read straight out of the app's own catalog and the credit cost taken from the same estimator the in-app picker uses. No model is described better than it is.
Narration voices.
The narration is the clock for everything after it, so it is generated first.
Scene images.
One still per scene, generated in bulk and timed to the narration.
Muse Image
The default scene model: 2K frames and a large reference budget, at a value-tier price.
3 credits per image
How Autotube uses itNano Banana 2.1
Google's newest image model: the same cast scene after scene, the largest reference budget on this site, at a value-tier price.
11 credits per image
How Autotube uses itGrok Imagine
xAI's value champion: consistent characters at a few credits an image, cheap enough for a long shot list.
6 credits per image
How Autotube uses itGPT Image 2
OpenAI's precision model for one-off images and reference-driven edits, rather than bulk scene coverage.
Priced from the request
How Autotube uses itGPT Image 2.5 Flare
1080p scene frames at a value price, and the 1080p-class model your own Replicate key can pay for.
4 credits per image
How Autotube uses it
Scene clips.
Short moving clips for the scenes a still cannot carry.
Veo 3.1
The best-looking clip model here, generated one shot at a time. The premium option for a cold open.
128 credits per second of clip
How Autotube uses itVeo 3.1 Fast
Veo motion across a whole scene list, at a little over a third of the flagship's per-second price.
48 credits per second of clip
How Autotube uses itGrok Imagine Video
The value clip model — cheap enough per second to use moving scenes liberally.
16 credits per second of clip
How Autotube uses it
Pick the model per scene, not per subscription.
Every plan reaches every model in this list, and the app shows the exact credit cost before it runs a thing.