Up next in 10
This Chinese Qwen3 Omni is INSANE First Test | The BEST Open Source Multimodal AI Yet
Sep 23, 2025
Get the full source code of application here:
Show More Show Less View Video Transcript
0:00
Uh hey guys, welcome to this video. So
0:02
in this video, I'll show you a
0:04
Chinese-based model which is Quen Quen
0:07
3. As you know, they have released a new
0:10
model in the series which is called as
0:12
Quen 3 Omni. If you go to their official
0:15
website chat.quen.ai
0:17
inside the dropdown, they have
0:19
introduced a new model in the series
0:21
which is Quen3-omnis-lash.
0:25
So this is the latest model that they
0:27
have added. I think they have added
0:30
yesterday. So once you uh change this
0:33
model here. So this supports ultimate
0:36
you can do all the things that you can
0:38
do with previous models but this is
0:41
slightly better. If you also read their
0:43
official blog Qu3 Omni.
0:47
So this is the latest model that they
0:49
have released. It's also available on
0:51
hugging face as well. So as you can see
0:54
this is the latest model Quen 3 Omni 30
0:57
billion parameter
0:59
and it's almost got 2,000 installed. So
1:02
you can even install this model inside
1:04
your own machine as well if you have a
1:06
powerful machine. So quite a number of
1:09
spaces are there.
1:11
You can try this specific model directly
1:14
on hugging face. So you can basically
1:18
load this model.
1:21
So it basically has let me just open it
1:25
in my account that I log in.
1:31
So simply go to this link on hugging
1:33
face to try out this model. It's an
1:35
ultimate model for audio video image
1:38
generation.
1:40
So I personally tried this. It's really
1:42
insane and it's really better as
1:44
compared to the previous models inside
1:46
Quen.
1:50
So it's basically just loading it. So
1:52
just wait.
1:54
So after it loads you can use this. So
1:57
basically you can see you simply write
1:59
the system prompt. It can even generate
2:03
voices as well. So I can simply say
2:09
I written the prompt here. After that I
2:11
click
2:14
you can even upload your own video as
2:16
well. This is audio video input.
2:29
You can interact with the model through
2:31
text, audio, images or video. Use the
2:33
tabs to
2:35
this is the offline mode right here. You
2:37
see?
2:44
Or you can simply go to quench
2:48
chat.quen.ai
2:50
their official website. Simply select
2:53
this model.
2:56
Simply select this model and after that
2:58
you can do all sort of things. You can
3:01
change the images.
3:06
You can generate videos. Image
3:08
generation. So I will say two
3:12
girls coding on laptop.
3:15
You can generate images, videos, coding
3:17
applications. All this is an insane
3:20
model coin 3 Omni flash. They have just
3:23
recently released this and either you
3:25
can try this on hugging face or you can
3:27
try this side your own official website.
3:31
It's completely free opensource and uh
3:36
can see it actually generated the image
3:39
for you.
3:42
You can even change the aspect ratio as
3:44
well. So this is the model guys.
3:47
Definitely try this
3:49
3 ohm flash.