---
title: 'CODING WAR: DeepSeek V3.1 vs ChatGPT-5 PRO - Which AI Codes Better? 🔥🔥🔥'
source: 'https://youtube.com/watch?v=jQj14_yovIY'
video_id: 'jQj14_yovIY'
date: 2026-06-17
duration_sec: 0
---

# CODING WAR: DeepSeek V3.1 vs ChatGPT-5 PRO - Which AI Codes Better? 🔥🔥🔥

> Source: [CODING WAR: DeepSeek V3.1 vs ChatGPT-5 PRO - Which AI Codes Better? 🔥🔥🔥](https://youtube.com/watch?v=jQj14_yovIY)

## Summary

This video pits DeepSeek V3.1 against GPT-5 Pro in a series of coding challenges, testing their ability to generate functional front-end projects. Over seven prompts, the creator evaluates each model's output quality, speed, and feature accuracy, ultimately declaring a winner based on a point system.

### Key Points

- **Challenge Setup** [0:00] — The video compares DeepSeek V3.1 Deepthink and GPT5 Pro in a coding battle, focusing solely on coding benchmarks.
- **Flappy Bird Clone** [0:43] — GPT-5 Pro's version worked perfectly with accurate collision detection and sound effects, while DeepSeek's had a missing pipe gap and faulty pixel detection.
- **Snake Game Challenge** [2:11] — DeepSeek's snake game worked but had bugs causing repeated game-over. GPT-5 Pro's version was completely non-functional, with buttons unresponsive and the game starting prematurely.
- **Portfolio Website** [3:38] — Both models produced good results, but the creator preferred GPT-5 Pro's neon theme and overall design, giving it the point.
- **Shopping Website** [5:35] — GPT-5 Pro's website was fully functional with working filtering, quick view, and add-to-cart. DeepSeek's version was mostly non-functional 'showpiece'.
- **Business Dashboard** [7:08] — DeepSeek excelled with a real-time updating UI, despite a minor scrolling bug. GPT-5 Pro's dashboard had a terrible UI and no working functions.
- **Crypto Tracking Website** [8:19] — DeepSeek outpaced GPT-5 Pro with a well-designed site featuring price alerts, portfolio, and candlestick charts. GPT-5 Pro's output was unfinished and disorganized.
- **3D Solar System** [9:24] — GPT-5 Pro's version had better visuals, clickable planets with info, and asteroid layers. DeepSeek's was decent but had non-working clickable planets and strange planet visuals.

### Conclusion

GPT-5 Pro narrowly wins the coding battle by one point, but DeepSeek V3.1 offers excellent value and competes closely, often excelling in specific tasks like real-time data dashboards. The creator suggests both models have strengths, but for this benchmark, GPT-5 Pro takes the win.

## Transcript

Hi guys, welcome back to the channel. In
today's video, we're going to show you a
comparison between Deepseek 3.1 versus
GPT 5 Pro model. In our previous video,
one of our subscribers commented asking
us to create a video on GPT Pro versus
Deepseek 3.1. So, here we are with
another brand new video. So, let's jump
straight into the action. Setting up our
coding battle.
So here as you can see we have the GPT5
Pro model and on the other hand we have
DeepSeek V3.1 Deepthink.
In this video we're purely going to
focus on coding benchmarks. So this is
going to be intense.
>> Here comes our first prompt. Generate a
complete Flappy Bird clone using HTML
CSS. include pixel perfect collision
detection, smooth animation, score
tracking, and sound effects using web
audio API. Let's give this prompt to
both models and let's see what happens.
So, as you can see, DeepSeek started his
thinking and GPT Pro is thinking, too.
And to my surprise, Deepseek finished
his thinking and also started writing
code.
Sometimes it's fast, too. That's
interesting. Okay, now Deepseek is
finished. Let's check the code and
Here's the game.
>> Let's click on start.
>> Guys, what the heck? How am I supposed
to cross this pipe if there's no gap?
>> And also, pixel detection is not
working. It's repeatedly showing game
over.
>> This is not good, Deep Seek. Now, let's
check GPT Pro. So, here's the finished
code. Let me copy the code and test it.
Let's click on start. And yeah, it's
working. It's perfectly working. The
pipe detection is working accurately.
Score function is working too. And it
has nice sound effects. Not harsh, not
too subtle.
>> So overall, it's working perfectly. One
point to GPT5 Pro.
[Music]
>> Test two, snake game challenge. Now,
here's our second prompt. Create a
modern snake game with neon glowing
effects, dynamic grid scaling, multiple
difficulty levels, food powerups, smooth
snake movement, and gameover animations.
Let's send this prompt to both GPT as
well as Deepseek model.
Okay, so here Deepseek is finished. Let
me copy this code and test it. Let's
click on start game. And it seems it's
not working. Let me click on easy level.
Okay, it started working. I think there
are some function misalignments, but
anyway, let's play this game. Okay, did
you guys notice why we got game over
H? Okay, I got it. We're getting the
same game over again and again. There
are some errors and bugs in the code.
So, DeepS works, but not fully
perfectly. It needs to fix some bugs and
then only it can run properly.
Now, let's check GPT 5 Pro. So, here's
the GPT output. Let me copy the code and
test it. We got game over before we even
started the game.
GPT, you really messed up. Terrible
mistake. No buttons are working and the
game already started before we even
pressed start.
>> Deepseek gets one point. At least DeepS
was working. Now both models have equal
points. Let's move to the next prompt.
Generate a cuttingedge portfolio website
for a freelance video editor featuring
video reels. Showcase with custom video
player, before after project
comparisons, client testimonials,
carousel, timeline of projects, contact
form with animation, and parallax
scrolling effects.
Let's send this prompt to each model.
And I'm noticing that today Deep Seek's
thinking is very fast. Like it's hardly
taking minutes and starting the code. So
that's a good point. There's definitely
improvement in coding. I can surely say
that about Deep Seek. Okay, so Deepseek
is finished. Let me copy the code and
test it. Here's the website created by
DeepSeek model. Looking good. The color
theme, animation effects, and it has all
the requirements that meet our
expectations.
There are client testimonials as we
asked for and project timeline as well.
So, it accurately followed our prompt so
well.
Good job, DeepSeek. Now, let's test GPT5
Pro.
Okay, copy the code and let's see the
output.
>> Okay, guys, first reaction. Wow, the
color theme picked by GPT is amazing.
Like this blue neon theme giving that
tech vibe like it should as this website
is designed for a video editor.
So, the color theme is excellent and
there are video samples which are empty,
but they're there. And the project
timeline, the project timeline looks
good to me. If you look closer, there's
slight misalignment. It should be at the
center of the line, but anyway. So,
there's a contact form with slight
animation. Overall, I like the GPT
website more than Deepseek. It could be
a personal choice, too. Some of you may
like the Deepseek website, but for me
personally, I'm liking the GPT website.
So, GPT gets one more point. Now, GPT is
leading in our comparison. Let's move to
the next prompt.
Test four. Create a high-end shopping
website with product filtering system,
interactive product viewer, shopping
cart with slide out animation, wish list
functionality, product zoom on hover,
rating system, and checkout process with
form validation. You can see we're
giving hard and specific functions so
that we can see who follows the prompt
more accurately.
Let's send this prompt to both models.
And you can notice DeepSk's thinking is
very fast today.
I noticed that in coding, especially
DeepSeek improved its thinking. Now it's
more efficient and faster than the
earlier version. Okay, now Deepseek is
finished. Let me copy the code and test
it. And here's the DeepSseek output. The
theme is very generic, very basic
website, and product filtering functions
are not working very well. None of the
buttons are working, so it's just a
showpiece, nothing else. DeepSeek
disappointed us here. Anyway, let's
check GPT5 Pro output. Okay, so here's
the code. Let me copy this and check.
And now this is the website generated by
GPT5 Pro. And I must say this looks
good. No joke. The filtering functions
are working perfectly. Quick view is
also working and add to cart is fully
functional.
>> So GPT really put some effort here and
made it functional as well as we wanted.
So one point to GPT again.
Now GPT is two points ahead of Deepseek.
Let's give our next prompt.
>> Test five. Generate a comprehensive
business dashboard showing AI
development metrics, profit analysis,
growth charts using chartjs, animated
counters, and real time data simulation.
Okay, this is a bit harder than the
earlier ones. Now let's send this prompt
to both models. While deepseek is
finished, let's test it.
Okay, this time DeepSeek excelled at
this. It's updating in real time and the
UI is looking good to me, but there's a
bug I noticed. It keeps updating in real
time, but there's no function to handle
long earlier data, which is making the
website length keep going on and on. But
overall, this website is looking good.
Now, let's test GPT 5 Pro output. And
here's the output. And GPT again messed
up this. The UI looks terrible with no
functions at all. There's nothing to
watch. Okay, let's give the point to
DeepSeek. Now, Deepseek is only one
point behind GPT. What do you think,
guys? Will Deepseek overtake GPT5 Pro
here? Comment down below your
predictions, and those who accurately
predict will get a shout out in the next
upcoming video. Test six, cryptotracking
website. Okay, so here's our next
prompt. Build a cryptotracking website
with real-time price updates,
interactive candlestick charts,
portfolio calculator, price alert
system, and animated price change
indicators with color coding. Let's send
this prompt to Deepseek as well as GPT5
Pro. Okay, Deepseek is finished. Let me
copy the code and run this. And here's
the output. It's looking good with all
the functions we expected. There are
price alerts and portfolio and
candlestick charts. And the UI is
overall looking good. Now let's test GPT
5 Pro. Okay, so copy this code and
here's the output.
GPT, what happened to you, boy? This is
what they call
>> Well, let's just say it looks very bad
and unorganized and it's not even
finished properly. Deep Seek again. Got
this point. Now Deepseek and GPT are at
the same point. Now let's give the final
prompt and we'll know who will be the
winner of this video. Deepseek 5 3.1 or
GPT 5 Pro. Final test 3D solar system.
Here's our last prompt. Generate an
interactive 3D solar system using CSS 3D
transforms in JavaScript. Include
realistic planet orbits, clickable
planets with information panels, zoom in
and zoom out functionality with planet
rotation animation. Let's give this
prompt to both models. Okay, Deep Seek
is finished. Let me copy the code and
test it. And here's the output.
The sun looks good, but the planets are
kind of strange.
We have zoom in andout features as well
as a reset button. And the planet
clickable function is not working, but
overall it's looking decent.
Now, let's test GPT5 Pro Code. Let me
copy the code. And here's the output.
This looks much better than DeepSeek for
sure. The sun looks good here. and
planet visuals are better than deepseek.
It also has zoom in andout features and
has asteroids in the outer layers. And
one more thing this has is the planet
clickable function which shows the
information of that planet when you
hover over it.
So guys, I think GPT5 Pro clearly gets
the point here. Final verdict and
conclusion, GPT5 Pro model wins this
test by one point. Deepseek is very
cost- effective and in terms of value
for money and effectiveness, Deepseek
will win. But in this pure coding test,
GPT won this time. Both models are great
at coding and sometimes Deepseek is
better than GPD5 Pro. But if you really
ask me who's the winner, then I would
say both are winners. But for today's
video, in terms of benchmarking, GPT5
Pro is the winner of this video. GPT5
Pro sometimes gives exceptional
front-end code that looks very good
compared to Deepseek's basic generic
websites that look very plain sometimes.
For the next video, do comment below
which model you want to test next. I'm
thinking of bringing Claude 4.1 Opus
thinking model and Kimmy model 2 for our
next videos. Do let me know your
thoughts. On this note, let's end this
video here. If you haven't subscribed
already, then subscribe to our channel.
Many exciting videos are coming very
soon. Thanks for watching and I'll see
you in the next one.
>> Through the wastelands, through the
highways, till my shadow turns the sun
raise. And on we
[Music]