Qwen Image | Game changer in AI Art

Quinn image is described as a gamecher

for AI art and general image generation.

But is it really? Today I'm putting it

to the ultimate test to see if it stacks

up to the expectations. We're going to

see how well it sticks to prompts, how

good its native text pretty is, how well

it generates the images of realistic

people, and whether it's worthy the

cost. Let's get started. To get started,

I'll choose create photos. Once we are

here, we can choose Quinn from these

number of models that are here. So just

scroll down and we have Quen image here.

Then for the image size, I'll choose

this landscape ratio so that we can see

it proper. Then to get started, we're

going to test the prompt lance of Quinn.

We want to see how well this model

follows instructions. So we're going to

do two tests for promptance. The first

one is this. I'll just paste it here. So

when you read through this prompt, I'm

sure you already have the picture. Now

we want to see if the picture that we

have in mind is similar to what the

model is going to generate. So once

you're okay with all your settings, we

can just hit generate.

Okay, we have the results there. And

this is very good. We can see that it

followed the prompt exactly. We had the

two tables here. The woman in the B

court, the bicycle in lead, the leaning

against the wall. And this is very good

except for this table. It doesn't have

the newspaper. But we can see that this

follows the prompt very very closely

which is very good. I'll give it a 9 out

of 10 for this specific image. Let's try

a different prompt. Okay. I've just

paste in a different prompt here. Again

I have my own imagination and I want to

see this will be very close to the image

that the AI will generate. Once you're

okay with everything we can click

generate again. We have the results

here. Again, this is spot on. Everything

that we described in the prompt is here.

This is very good adherance to the

prompt. I'd actually give this one a 10

out of 10 because everything I've said

in the prompt, I can see it here. This

is very good. So, I can say Quinn passes

the prompt test here. It follows

everything very well. Let's move on to

text. So, Quinn is built to excel at

text. It's one of the selling point of

Quinn. So let's give it a try here.

We'll try to see if it lives up to the

expectation. So here I'm asking you to

design a wedding card which is

textheavy. So we're testing for design

and we are specifically interested in

the text that it's going to generate. So

this is the prompt that I have here. By

the way, if you're interested in all

this prompt that I've used, I'll share

them in the description below. So we

have our prompt here. Let's hit generate

to see what will happen.

Wow, look at that. This is brilliant.

This is good. Except for this one here.

You can see it's not readable, but the

rest is good. Exactly what we gave it

and it has done the design. This is

excellent. So, if you have no design

skills and you want to make a wedding

card, invitation card or something like

that, I think this is a good model to

pick up. So, this is exactly what we

prompted for except for this one here.

So, I'd give it a 9 out of 10. Very good

results here. Many image generating

models would struggle with this. So,

let's try a different prompt, but we are

still testing how well it generates

text. So, here I've tested the prompt.

You can see there's a lot going on here.

The text we're interested in is this one

here and also this. Not just the text,

but how it's rendered. So, this one

should be written on a wooden sign,

while this one should be written on a

chalkboard and it should be handwritten.

So, let's see how it does with this one.

Okay, we have the results. This is good

except that this is not written on the

top board. So, the board is here. We

expected this to have been written here,

but this is okay. Everything else was

followed. I'll say this is an 8 out of

10. So, we can surely see that Quinn

does very well with text. So, let's move

on to see how it works with the images

of people. Now, we have this prompt here

and we expect to see three people in the

image. So, each of the person is doing

something different. And so we can see

one person is wearing sunglasses and

blasting. The second person is glinging

with their hands while speaking and the

third person is drinking from a glass of

iced tea. So many models struggle with

fingers and also the image as a lot of

people. Let's see how Quinn does here.

So just hit generate. And we have the

results here. Wonderful. This is good.

Let's look at the fingers. The fingers

are fine. I think everything is okay

here. In my mind, the picture I was

having is that all of them will be men.

But this is very good. I'll take nothing

away from it. I'll give it a 10 out of

10. Let's try a different prompt. So,

another thing that AI struggles with is

when someone is in motion or in action

of some sort. It may be running,

jumping, or anything like that. So,

that's what we want to test here. So

far, Quinn has proved to be very good

for image generation. Let's see how it

knows this one. We have the results

here. This is very very close to what we

asked for. There's only one problem. You

can see the running doesn't look very

realistic. I think even at first glance,

you can say that this image is AI

generated. And for that, I'll give it an

8 out of 10. Everything else is spot on

and it's very good. Lastly, let's look

at the cost or the pricing of this

model. So, I'll go back to the models

here. And by comparison, we can see that

GPT1 here needs 30 credits. tracks needs

20 10 we have two here let's scroll down

to Quinn and Quinn needs only five so

when we're comparing it to other models

here you can see it's one of the

cheapest it's the same with imagin only

this version of flax is cheaper than

quen so to be able to get all these

features for this course I think it's a

very good deal I would highly recommend

Quinn to any person who is interested in

air image generation you get good text

rendering good prompted balance, good

images of people, and the cost is very

low. So, let me know what your thoughts

are on Quinn image in the comment

section below. And if you want to see

how you can make lip syncing videos with

AI, that video is up here. I'll see you

over there.