# Opensource AI video in 2025 vs 2026 is wild! — retained record

> Retained reddit record from reddit:StableDiffusion.

Canonical: https://highsignal.app/data/records/f92c632500206f4b

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:StableDiffusion)  published 2026-09-19 16:57 ingested 2026-09-20 02:37

[original source ↗](https://www.reddit.com/r/StableDiffusion/comments/1wkrcfb/opensource_ai_video_in_2025_vs_2026_is_wild/)

## retained text



```
It’s crazy how much changed in just one year.  
.  
I put together a 2025 vs 2026 comparison, and the jump in video quality, motion, consistency, and overall realism is honestly massive.  
For the 2025 video, I had to use DaVinci Resolve quite a bit to edit and fix things, sounds. With the 2026 video, I didn’t need DaVinci at all - except to put the comparison itself together.
```



document fetched 2026-09-20 00:30 · canonical https://reddit.com/r/StableDiffusion/comments/1wkrcfb/opensource_ai_video_in_2025_vs_2026_is_wild
