@vib3coded Yeah, the difference is definitely noticeable in this comparison. Fable captured the atmosphere and details of the scene better, especially in this style. It would be interesting to see a few more comparisons like this
Okay, this comparison is actually embarrassing😂
Left: Cloud Fable 5.1 - $2
Right: ChatGPT-5.6 Sol - $1
And yeah, Sol is cheaper. But apparently it decided to save money on the output too💀
Fable absolutely cooked here. It actually nailed the whole scene - the little house, Christmas trees, snowman, reindeer, snow, and that cozy pixel-art snow globe vibe. Everything fits together and looks intentional
Then you look at Sol and… bro, what happened? 😭 The composition is all over the place, the snowman is basically gone, the snow turned into one giant white blob, and the whole thing just feels unfinished
And that's what makes this comparison even funnier. Fable costs twice as much, yet the difference in quality is pretty hard to ignore
Looks like the washing machine has a new owner🧺🤖
This robot just grabs clothes from the laundry basket and loads the washing machine
Sounds pretty boring, right? That’s actually the impressive part
Teaching a robot to move its arms is one thing. Teaching it to find clothes, pick them up, understand the space around it, and carefully place everything into the drum is a completely different level
And honestly, everyday chores like this might be the real test for humanoid robots
Today - loading the washing machine.
Tomorrow - cleaning your entire apartment?
What’s the first household chore you’d happily hand over to a robot?👀
8.64 seconds... Another record broken. 🔥🔥🔥
The 100 m large-size final at the second World Humanoid Robot Games.
TianGong (Tien Kung) Ultra took gold in 8.64, successfully defending its title and beating the human world-record time again.
TianGong swept gold and silver.
HOLY SHIT, ASTRA JUST DROPPED🚀
The rumors were real-it literally just went live! Is the CBT actually up and running for you guys right now, or is it totally fried?
I’m literally dying to run some benchmarks and see what this thing can actually do under the hood
Previously, I brought a static koi pond image to life using ChatGPT-5.6 Sol, Grok 4.6 Build, Claude Opus 5, and GLM-5.3. This time, I decided to try the same idea with Gemini 3.8 Flash🐟
The goal remained the same: turn a still image into an interactive WebGL scene with a 3:1 aspect ratio and water that feels alive
What do you think of the result?👀
Is Astra coming out today? I have no idea🤷♂️
Meanwhile, I’m continuing to test Claud Fable 5.1 and Gemini 3.8 Flash
This time - a sailing ship on the open sea at sunset. And the difference really stands out
Look at Fable’s version: the rigging, translucent sails, lanterns on the stern, reflections, and that shift from gold to purple across the water. It all comes together into a scene I could just leave on my screen and watch
I especially love how the light passes through the sails. Those little details are what make the scene feel convincing to me
Gemini produced a noticeably simpler version: less detail on the ship, simpler materials and lighting. The waves and movement are there, but it doesn’t give me the same sense of atmosphere
That said, I’m using Gemini for free, and that’s an important part of this comparison. Getting a scene like this for $0 is still pretty cool. But visually, Fable 5.1 is comfortably ahead for me here
Astra, if you do launch today, I’ve already got a challenge waiting for you👀
WAIT, AND THIS IS FREE?🤯
Gemini 3.8 Flash just launched, and I’ve already tested what kind of ship in a bottle it could build. The result cost me $0
Inside the glass is an entire miniature scene: a three-masted sailing ship, voxel waves, a palm island, clouds, and a sun. I especially like the depth of the water - beneath the ship isn’t just a flat blue surface, but an ocean built in layers
And this isn’t an AI-generated video. You’re watching a recording of an interactive 3D scene: you can rotate the bottle, zoom in, follow the ship with the camera, and switch between daylight, sunset, and night
Gemini brought its own take to the idea - bright and playful, like a collectible diorama on a wooden stand
After experimenting with paid models, getting this result for free feels particularly satisfying
What do you think? Can it hold its own against the models from my previous comparison?👀
THIS IS INSANELY BEAUTIFUL. Just look at this comparison: Claude Fable 5.1 vs ChatGPT-5.6 Sol🤯
I gave both models the same challenge - create a living, interactive voxel world inside a glass bottle
Honestly, Fable 5.1 delivered an incredible result. Its scene contains around 1,700 animated ocean columns, a ship, a lighthouse, a palm island, the remains of an old shipwreck, seagulls, jumping fish, rain, lightning, and even a giant sea serpent appearing during the storm
Everything stays perfectly inside the bottle - nothing clips through the glass. The level of detail, atmosphere, and animation makes it feel like a finished game scene
The cost:
• Claude Fable 5.1 - $4.60 • ChatGPT-5.6 Sol - $3.60
Fable went all-in on detail and cinematic atmosphere, while Sol created a completely different interpretation of the same idea for slightly less
Which one would you choose?
GPT-5.6 Sol just made Fable 5.1 look wildly overpriced
We gave both models the exact same challenge: generate five connected Three.js scenes of a futuristic island mansion-zero retries, zero manual fixes
The bill?
🔴Fable 5.1: $5.69
🟢GPT-5.6 Sol: $0.88
Fable delivered richer visuals: realistic waves, deeper environments, thousands of chandelier crystals, a massive water-slide tower, and a smoother helicopter landing
But it also forgot one minor detail…
Water. In the pool
Sol’s result was simpler and more stylized, yet consistent across every scene - and its tiny water park actually worked
So yes, Fable gives you more detail
It also costs over 6× more and still makes hilarious one-shot mistakes
Would you pay the premium - or take the cleaner $0.88 result?
WHAT IS FABLE 5.1 EVEN DOING?! JUST LOOK AT THIS RESULT🤯
It literally came out today, so I immediately decided to see what it could do
I asked it to create an interactive Japanese-style scene. Instead of giving me some basic animation, it built an entire floating city directly in the browser
A massive multi-level pagoda, dozens of buildings, blooming cherry trees, gardens, pathways, and torii gates - all fully 3D
You can rotate the entire scene, zoom in, and explore it from any angle. And no matter where you move the camera, you keep discovering more details
I genuinely didn’t expect to see something this huge and polished on the very first day of the model’s release
If Fable 5.1 can already do this at launch, things are about to get very interesting🔥
Claude Opus 5 built this school of manta rays in p5.js
It’s a single HTML file with no textures or premade models. All the geometry is recalculated from scratch every frame
The wings are built along their own arc length, so they don’t stretch while flapping. A traveling wave moves across the wings while each chord twists around the spar - that’s what generates thrust
The tail isn’t manually animated either. The body stores its path from the last 130 frames, and each tail segment follows a previous direction
Small details make everything feel alive: banking during turns, vertical body movement, and cephalic lobes that unfold at higher speeds
The size of each ray represents depth. Distant rays become smaller, blend into the water, and are rendered first
There’s also a Show Forces mode. Set Wing Flexibility to zero, and the twist disappears. The ray keeps flapping but barely moves forward
All that speed was hiding in a single angle
Fable 5.1 is expected to drop today, and I’ll definitely be putting it to the test