This is an AI video generation comparison for image-to-video 
prompt:
Animate this pencil drawing of a cute surreal magical adorable scene, let penguin walk turning around itself while chasing butterflies. Make sure all the butterflies are animated.
Tested: July 30, 2025
This was made a while ago, actually. Had to instruct it to animate all butterflies because some would remain frozen for whatever reason.
Tested: July 31, 2025
The penguin is a bit hyperactive and butterflies sleepy.
Tested: July 31, 2025
Very cute animation is done by the full 14 billion Wan 2.2 model.
Tested: July 31, 2025
Vidu can animate these for sure. I like how his eyes blink. Might've been even better with more movement though.
Tested: July 31, 2025
The Lite version is still quite capable, and in this case worked well with this simple prompt.
Tested: July 31, 2025
The full model might even have less exciting motion than Lite version, but digital illustration style details might be better here.
Tested: July 31, 2025
Veo 2 seems to struggle with this task. Generation 2/2, both have identical issues
This task tests how a model is able to work with digital illustration media. In this specific image, there is a hand-drawn penguin and a few colorful butterflies on solid white background, so we just want to see if AI is able to make them move in a believable, artistic and cute fashion. Common issues include jittery motion, lack of motion, partial animation (not all butterflies) - not sure why, they all seem clearly defined, not blurred, colorful, so should be easy to spot, but because this has been a known issue before I've included an additional direction to focus AI's attention on spotting all the butterflies.
Check out the results from Hailuo AI x Minimax (Hailuo 01 Live) vs Wan (Open-source) (Wan2.2 TI2V-5B) vs Wan (Online Platform) (Wan 2.2) vs Vidu AI (Vidu Q1) vs Freepik (Seedance 1.0 Lite) vs Freepik (Seedance 1.0 Pro) vs Veo (Veo 2) for similar or identical prompts side-by-side.
