This is what 12gb of vram builds in 2026, absolute magic
Rtx 3060 12gb, #1 gpu on steam bonsai 2 27b + mtp, 5.95 gb of weights hermes agent, 5 hours, 328k tokens written 8 js files, 2,368 lines, zero hand…
Wednesday afternoon AI NEWS roundup 🤖 23.09
🎙️ Alibaba releases the Qwen-Audio-3.1 audio suite with new TTS-Next and ASR-Next models - 📱 Alibaba unveils Qwen Intelligence: a three-agent AI…
Side by side comparison ThinkingCap: Qwen 3.8 27b vs. Original model
Esto es lo que se logro con MiMo v2.6 Pro
Creo que es un muy buen resultado. Pero demoro alrededor de una hora en hacerlo y otra media en grabar el video.
In molti mi avete scritto: «ma quale modello posso scaricare con Ollama e far girare sul mio computer?»
Modelli come GLM 5.3, Qwen 3.8 o Kimi K3 hanno una qualità paragonabile a quella dei modelli di frontiera. Solo che sul computer non partono.
This Week AI Is Insane - HUGE AI News
Claude Opus 5.5 - GPT-6 Soul & Luna - MiMo V2.6 - Qwen 3.8 Omni Flash - Jing and Dao - Qwen Image 2.1 - Viggle Meridian - Ming Image 0.1 Design -…
Last night after fixing Qwen 3.8 Flash Next to get a better return on tok/s on decode while retaining a…
Last night after fixing Qwen 3.8 Flash Next to get a better return on tok/s on decode while retaining a coherent state it worked flawless over…
Who do you think created this Empire State Building model?
Opus 5.5? GPT 6 sol? Grok 4.7? Qwen 3.8?
I’ve spent a long time experimenting with my own quantization of Qwen 3.8 Flash and a custom harness.
The results are starting to feel really good. Here’s a little cozy pixel-art game about feeding fish in a pond, made from this prompt: “Make a small…
Cloudfare worker AI gives 10,000 free neurons every single day😳🔥
Cloudflare Workers AI gives you 10,000 free neurons daily basically free compute to experiment with powerful models.
I handed Qoder a real task and watched its AI agent turn a goal into a finished result.
I chose the workspace, model, and permissions, then watched Qoder plan, use tools, make changes, and verify the output, all from the new task screen.
I just gave qoder_ai_ide a real task: build a functional AI project management dashboard from scratch.
I described the outcome, chose Qwen 3.8 Flash, set the permissions, and watched it plan, build, test, and fix the project.
More FREE AI models just dropped on the board
DeepSeek V4.1 Flash, GLM 5.3 Flash, Grok 4.7. all FREE: Freebuff: DeepSeek V4.1 Flash, GLM 5.3 Flash, MiMo 2.6 Flash, Solar Pro 4, Muse Spark 1.2:…
My laptop couldn’t work with WiFi onboard of long flight
So I used junie_ai Local with Qwen 3.8/3.6 Blend and it helped me to troubleshoot and connect properly 😜
Qwen 3.8-FLASH is free right now
We can use Alibaba’s new 125B model for free through Qoder until September 30.
Included UD-Q3 and UD-IQ4 in my Qwen 3.8 27B quant benchmark.
Feed the model a task as an image card, then see if it can produce the animation.
Qwen 3.8 at 51tps on Studio M5 Ultra via MLX-VLM ordering me sushi for dinner via computer use.
(Yes I know there is a DoorDash cli - this is an eval I’m called SushiBench)
SITUATION EXPLAINED: Alibaba published its AI roadmap.
Qwen 5 will be 5 to 10 trillion parameters. • Qwen 4 is in training now, with Qwen 4.5 and Qwen 5 scaling to 5 to 10 trillion, roughly the size of…
I had Opus 5.5 build a tiny Sims town where every resident's dialogue is written live by a local AI 🏘️🧠
Then I became the sky ☁️ and dropped a tornado on them 🌪️ and abducted the mayor with a UFO 🛸👽 Now they gossip about it 😂 Running 100% local on…