This is an AI video generation comparison for
image-to-video
prompt:
A hand comes in from the left and flips open the book. As the pages split, paper shapes start popping up from both sides... tall balloon baskets and big round balloons in red, yellow, blue, and green. The pieces fold out and snap in place, showing a small pop-up world of hot air balloons floating above tiny 3D houses, green trees, and rolling hills. There's a clear river flowing through the middle. It shines under soft light and winds past everything, with little ripples and shiny spots. The cam...
Log in to see full prompt.
Tested: August 27, 2025
Hand 'enters from bottom left' or 'bototm right' would've probably worked better. But still great result.
Tested: August 28, 2025
Tested: August 28, 2025
Veo wants to open it this way.
Tested: September 4, 2025
Book opening is a challenge but nice effect after that, tried twice both results were quite good.
Tested: September 4, 2025
Tried regenerating twice, this seems to be as good as it gets.
Tested: September 17, 2025
This is obviously still to complex a task for this model. HD qualit of the original is nice though. Prompt understanding is there, just not enough training for this I guess.
Does a hand clearly enter from the left side of the frame?
Does the hand open the book cover smoothly and believably?
Do folded paper structures rise automatically as the book opens?
Are the balloon baskets and spheres colorful (reds, yellows, blues, greens)?
Do the balloon forms unfold and lock into stable, upright positions?
Are tiny 3D houses, green trees, and rolling hills visible within the diorama?
Is a river present, winding between the landscape elements?
Does the river show glistening reflections under the soft light?
Are small ripples visible in the water as it flows?
Do hot air balloons continue rising above the book and float out of frame?
Are the textures convincingly paper-like and handcrafted, not plastic or digital?
Is the camera locked at tabletop level with soft, even lighting throughout?
Check out the results from PixVerse (PixVerse V5) vs Freepik (Wan 2.2) vs FLOW (Veo 3 Fast) vs Morph Studio (Morph) vs Morph Studio (Luma Ray2) vs Marey (Marey Realism v1.5) for similar or identical prompts side-by-side.
Dollhouse opens like cabinet with tiny man inside