From being totally shook by Sora, to JI Meng fighting back—domestic AI video has been pretty wild these past two years.

I saw that “counter-kill” take floating around, and honestly, as someone who edits videos with these tools every day, it really hits home.

I still remember when Sora first dropped—everyone in the scene was panicking, thinking the gap between domestic and foreign tools had suddenly blown wide open. But looking back over the past year or two, the domestic side has caught up way faster than expected. Tools like Jimeng actually feel smoother for daily work—better output efficiency, way better Chinese semantic understanding.

Sure, “counter-kill” is a bit of a strong word, and who’s actually better depends on what kind of work you’re doing. But the shift from just playing catch-up to actually going toe-to-toe? That alone shows the gap wasn’t as scary as everyone thought back then. What’s your main tool right now? Let’s talk about the weak points of each.

I was really panicking at first, but now I’m way more chill about it.

It’s pretty obvious that domestic models understand Chinese prompts way better — the difference in experience is really noticeable.

“Counter-kill” might be a bit of an exaggeration, but it’s definitely good enough for everyday image generation.

Honestly, every tool has its weak spots — none of them have nailed character consistency yet.

Man, the pace of catching up is insane. Both closed-source and open-source are going all out.