The video directly compares two popular AI models using a practical prompt, offering a clear demonstration of their outputs and a subjective but well-reasoned recommendation. This format appeals to users interested in AI tools and development.
Summary
The video compares two AI models, Fable 5 by Anthropic and GPT-5.6 Sol by OpenAI, by giving them the same prompt to build an app that turns procrastination into progress. Fable 5 is shown to have better functionality with gamified elements and a timer, while GPT-5.6 Sol offers a more professional UI. The creator ultimately prefers Fable 5 for its overall features.
Structure
1Introduction of Fable 5 and GPT-5.6 Sol
2Prompt given to both models
3Fable 5's app output and features
Original caption
Testing Fable 5 vs 5.6 Sol inside of the Cosyra app. Which model is better? #vibecoding #mobiledev #aiagents #terminal #cosyra
More from @trycosyra
More breakout videos from this creator.
4GPT-5.6 Sol's app output and features
5Comparison of the two models
6Creator's final recommendation
Product placement
Anthropic Fable 5 (AI model, generates text, used for coding tasks)
OpenAI GPT-5.6 Sol (AI model, generates text, used for coding tasks)
Cosyra (App, allows users to run code and AI models on their phone, used to test AI models)
Nudge (App, helps users break down tasks into smaller steps, gamifies productivity)
On-screen text
Fable 5
by Anthropic
ANTHROP\C
5.6 Sol
by OpenAI
OpenAI
Cosyra
Search sessions
Terminals
Chats
Running apps
node
node
Playwright Live
Playwright Live
Open sessions
Fable 5 test
GPT-5.6 Sol test
Recent
New chat
New chat
Archives
Casey Media
caseymediaofficial@gmail.com
model:
gpt-5.6-sol-high
/directory:
Tips: Our most capable model yet, GPT-5.6 Sol can tackle common code changes, dig into research, and perform complex tasks. But it's also capable of producing polished documents, and take on lower reasoning efforts--try starting lower, then work your way up for harder jobs.
You have 4 usage limit resets available. Run /usage to use one.
Model changed to gpt-5.6-sol-high
Build an app that turns procrastination into progress. Optimize for mobile. Call the project Fable-test. Run it on the Local host when the project is complete.
Starting ROS servers (1/2): codex_apps
/n review on my current changes
gpt-5.6-sol-high
Nudge
YOUR GENTLE RESET
Good afternoon.
Let's make it lighter.
You don't need more motivation. You just need a smaller first step.
What are you putting off?
e.g. Finish my project...
We'll turn it into a 5-minute win.
MOMENTUM
Small wins add up
Today
36
SCORE
Nice momentum
You're building a steady rhythm today.
4 days
gentle streak
0 wins
completed today
YOUR NEXT MOVES
Today's tiny wins
0/3
Open the project and write one sentence
Done - look at you go
Reply with a quick "got it"
Done - look at you go
Put the documents in one folder
From: Sort out my tax return
+ Add another tiny win
Done is allowed to be small.
Progress counts even when it's messy, brief, or invisible to everyone else.
Today
Wins
Focus
You
54
SCORE
Nice momentum
You're building a steady rhythm today.
4 days
gentle streak
1 wins
completed today
YOUR NEXT MOVES
Today's tiny wins
1/3
Open the project and write one sentence
Done - look at you go
Reply with a quick "got it"
Done - look at you go
Put the documents in one folder
From: Sort out my tax return
+ Add another tiny win
Done is allowed to be small.
Progress counts even when it's messy, brief, or invisible to everyone else.
Today
Wins
Focus
You
90
SCORE
Nice momentum
You're building a steady rhythm today.
4 days
gentle streak
3 wins
completed today
YOUR NEXT MOVES
Today's tiny wins
3/3
Open the project and write one sentence
Done - look at you go
Reply with a quick "got it"
Done - look at you go
Put the documents in one folder
From: Sort out my tax return
+ Add another tiny win
Done is allowed to be small.
Progress counts even when it's messy, brief, or invisible to everyone else.
Today
Wins
Focus
You
GENTLE FOCUS
Just this one thing
Choose one small thing and stay with it.
05:00
ready when you are
Start focus
I need a break instead
05:00
you're doing it
Pause
I need a break instead
04:58
you're doing it
Pause
I need a break instead
04:57
you're doing it
Pause
I need a break instead
Receipts
proof you're not who your inner critic says
streak
1 day
reclaimed
30m
today
9 wins
Last 7 days
minutes reclaimed
30m
Fr
Level progress
Lv 3 - Dabbler
Lv 4 - Starter
435 XP - 65 more to level up
Recent wins
Take out the trash
just now
5-minute kitchen counter sweep
just now
Start one load of laundry
just now
Clear just the coffee table
just now
Write the closing paragraph badly
just now
Drop in the numbers, no polish
just now
Start completely fresh
Today
Wins
Focus
You
Transcript
Fable 5 or GPT-5.6 Sol which model is better? I'm going to give the same prompt to each model and see which one is better at completing the task. The prompt that we're going to give them is build an app that turns procrastination into progress. Optimize for mobile. Call the project Fable-test. Run it on the Local host when the project is complete. Okay, we'll go ahead and put these prompts into the models, let them run and see what we get. Fable 5 got done in about 15 minutes. Let's go ahead and open it up and see what we got. It looks like it chose a darker theme which we didn't specify, but it went with the darker theme and more of a gamified vibe. So every time that you complete something, it gives you XP and then at the top you can see there's levels. So you start at level one and each time that you check something off, you get more points. It looks like at the bottom there's different tabs so this one is a timer, so that you can set a timer and focus on a specific task. So that's a nice addition. It looks like it also added a stats screen, which is pretty cool. You can see everything that you completed, you can see what days that you completed it on and how much time you reclaimed and today's wins. So I actually really like this analytics page. It looks like it also has the tabs at the bottom of the screen, so the wins here, the focus and the you tab, those both don't do anything, so functionality wise it is lacking a little bit. If I had to give it to one model, I would probably say Fable 5 just because I think that the functionality was a lot better with Fable 5. Now I will say that GPT-5.6 Sol, I think had a better UI keeping it professional, but obviously that is just preference. So I do think I have to give it to Fable 5. Let me know which model you guys think did a better job down in the comments. Okay, let's go ahead and see what GPT-5.6 Sol came up with. So it looks like it went for a lighter theme and more of a professional theme. It's got a score here for your momentum, and then obviously you can check off the specific tasks that you needed to get done for the day. So all of this looks pretty good. I actually quite prefer the professional UI instead of the gamified UI. So the timer here, it actually pulls up just like the Fable one, but it does have a timer and then obviously you can pause it. So I think the timer actually looks a lot better than the Fable one. So the timer, they got correct and then obviously at the bottom you've got the profile tabs. So those both don't do anything. So functionality wise, it is lacking a little bit. So I would say that GPT-5.6 Sol is probably better for a professional UI, but Fable 5 is better for functionality.
