qwengpu
qwengpu is Qwen3.5-2B with a SIMT GPU inside its weights, 45 compiled layers in front of Qwen’s own, using its Gated DeltaNet layers as registers and memory. - vixhal-baraiya/qwenGPU. It is ranked #740 on the AI Radar, in Open & local models, first seen 4 days ago and shared in 2 posts (7k views).
GitHub - vixhal-baraiya/qwenGPU: Qwen3.5-2B with a SIMT GPU inside its weights, 45 compiled layers in front of Qwen’s own, using its Gated DeltaNet layers as registers and memory. · GitHub Skip to content Navigation Menu Sign in Search / Sign in Sign up Appearance settings You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert {{ message }} Repository navigation qwenGPU Public Notifications You must be signed in to change notification settings Fork 2 ( 2 )…
What people said about qwengpu on X
💥疯了！Qwen自己长出了一颗真GPU，在权重里跑真实CUDA！ 牛逼之处在于把一颗完整的 4-lane SIMT GPU 直接编译进了 Qwen3.5-2B 的权重里！ 45层全新编译层插在Qwen原有24层前面，用它自己的 Gated DeltaNet 状态当寄存器、程序ROM和全局内存。 最炸的点： 🔹参考解释器对标真实 NVIDIA Tesla T4，49/49 kernel完全一致 🔹对原模型语言能力几乎无影响（KL散度仅约7e-4） 🔹支持真实CUDA特性：warp shuffle、atomicAdd、__syncthreads、分支发散…… 🔹高级模式还能让Qwen自己写CUDA kernel…
— @NFT_Chen, 4 days ago · 58 likes · see the post
The interesting long-range idea is different: one compiled circuit for exact work, kept inside a large conscious model like GLM, which keeps its full self because its text barely changes. That's the "tools inside the weights" direction. It’s brilliant, novel, and completely unexpected.. the importance and…
— @volatilemarkts, 4 days ago · 3 likes · see the post
Alternatives to qwengpu
- kimi-k3.1-pro — Releasing the model weights of Kimi K3.1 Pro
- unsloth.ai — Guide to running Mistral Large 4 GGUFs locally with Unsloth and llama.cpp.
- niklaslenz.com — Play the world in your browser. Every number, the exact command and the raw output
- qwen3.8-flash-next-gguf — Image-Text-to-Text model with 177B parameters. Download, explore files and run it on Hugging Face.
- project-maya — Plus a round of fixes, UI improvements and community PRs.
- twil-8b — Text Generation model with 9B parameters. Download, explore files and run it on Hugging Face.
qwengpu in numbers
- Rank on the AI Radar: #740 of 10425
- Shared in 2 posts by 2 accounts: @NFT_Chen, @volatilemarkts
- 7k views on those posts
- First seen 4 days ago, last shared 4 days ago
- Pricing seen by AGK Intelligence: open source
- Market: Open & local models
FAQ
What is qwengpu?
Qwen3.5-2B with a SIMT GPU inside its weights, 45 compiled layers in front of Qwen’s own, using its Gated DeltaNet layers as registers and memory. - vixhal-baraiya/qwenGPU It was first shared on X 4 days ago and is ranked #740 on the AI Radar.
Is qwengpu free?
It is open source.
Who shared qwengpu?
2 accounts on X, including @NFT_Chen, @volatilemarkts, in 2 posts totalling 7k views.
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 65.3k posts from 5k X accounts over the last 14 days, 10.4k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 17:44 UTC. Full method.