Kling 3.0 rolled out and shocked the market. It is surprisingly good, but side-by-side testing shows it still has some flaws.
Here is the result.
VEO 3.1
❌ Shaky camera (did not carry as usual the camera movement is the weakness of VEO)
✅ Brushes her hair back (Action is the strength of VEO)
✅ Lip-sync
✅ Zooms into the product
🆗 Emotion
Kling 3.0
✅ Shaky camera
❌ Brushes her hair back (not able to follow the prompt)
✅ Lip-sync
✅ Zooms into the product
✅ Emotion (Better than VEO personally think)
Overall, VEO 3.1 remains stronger at following fine-grained action instructions, while Kling 3.0 stands out in emotion and camera movement, but still struggles with prompt accuracy in certain cases.
Video demo: