Qwen 3.8 Max overhead wasteland destruction derby.
I've been loving this 1-bit dithered style lately. Qwen 3.8 max is going to be a nice win for everyone tomorrow I'm excited to see what size people…
I ran Qwen 3.8, Fable 5, GPT-5.6 and Kimi K3 through 23 identical real-world builds.
Games. Landing pages. Physics simulations. Web experiences. The winner changed constantly. And that's the useful part.
Meta just released Muse Spark 1.2 + Muse Code.
So I tested it against GPT-5.6 Sol, Qwen 3.8 and Fable 5. The benchmark scores aren’t the interesting part.
Kimi K3 took ~500MB of 3D assets down to 17.8MB
Then generated 56 historical images and optimized those too. Kimi K3 is no doubt leading DeepSeek V4, Qwen 3.8 Max, Opus 5, and GPT-5.6 Over 5 hours…
Alibaba's Qwen 3.8 Max is here with 2.4T parameters, claiming top spots in coding & multimodal tasks.
A 27B sibling offers near-95% performance for realistic business hardware.
Alibaba's Qwen 3.8 Max is here with 2.4T parameters, claiming top spots in coding & multimodal tasks.
A 27B sibling offers near-95% performance for realistic business hardware.
Alibaba's Qwen 3.8 Max is here with 2.4T parameters, claiming top spots in coding & multimodal tasks.
A 27B sibling offers near-95% performance for realistic business hardware.
There's a mystery model on Arena called Kiana and everyone has already decided it's Qwen.
I don't buy it. Look at how Moonshot names its stealth checkpoints. K2.5 ran as Kiwido. K2.6 as Kiwire. K3 as Kivine.
Qwen 3.8 max makes beautiful websites and components.
If Qwen 3.8 27b is close to as good we will all be incredibly lucky.
Qwen 3.8 just beat fable 5 and GPT-5.6 on some of my real-world tests.
But the bigger surprise was where each model completely fell apart. I ran 23 identical builds across Qwen 3.8, Fable 5, Kimi K3, and GPT-5.6.
Acabou de surgir um modelo de IA chinês, o Qwen Kiana
Que já está batendo de frente com gigantes como o GPT-5.6 Sol e o Opus 5 em testes preliminares.
I turned my AI tools into a 24/7 team that works while i sleep.
The biggest upgrade wasn’t a better model. It was giving every model the same desk and memory.
実際ローカルLLMってよく聞くけど、どうなの？という話を整理と収録してみた。
気になっている、という方は是非見てほしい。 Museが出てきて、GB10も二台になったので、OpenRouter経由と、ローカルLLMってどこまでのスピードなの？どの程度のものなの？という点について比較。
Qwen 3.8 Max: An open model just beat GPT and Claude at using a computer.
OSWorld tests if an AI can click buttons and finish real desktop tasks like a person.
Ok skalskip92 didn't lie, Qwen 3.8 Max has impressive vision capabilities!
The few-shot image prompting is amazing: you can annotate an image with one bounding box, and it will find the remaining objects 🤯 I tried it out on…
Qwen 3.8: Alibaba just dropped a 2.4 trillion parameter AI.
And you can run it yourself. It's one of the biggest models ever released. Open. Not locked behind one company.
Qwen 3.8 Max Full COURSE 1 HOUR (Build & Automate Anything)
However 3d modeling Qwen 27b BF16 just nails it.
Muse puts up a good show but it's not really close here. However like always Qwen uses way too many tokens in general, I wonder how 3.8 is going to…