GitHub radar
A local AI model that answers almost twice as fast
This is a ready-made AI model you can run on your own computer — it thinks faster than the standard version and doesn't need the cloud.
It's a fine-tuned version of Alibaba's Qwen3.8-27B model that runs on your computer, offline, and answers faster than the original.
When it helps
Useful if you want a local assistant for simple tasks and care about privacy or cost; don't expect it to match top cloud models.
Pros
- Answers almost twice as fast as the standard version
- Uses 58% fewer reasoning tokens (text chunks) on average
- Runs locally — no internet connection, no per-request cost
- Understands both text and images
Cons
- Needs strong hardware: a Mac with 32GB+ RAM or a 24GB GPU
- It's a mid-tier model — won't match top cloud models
- Commercial use above $1M revenue needs a separate license
How to set it up — step by step
- 1Open your AI agent (Claude Code, Codex) inside your project folder.
- 2Send the agent the text from the block below.
- 3Let the agent install Ollama (a tool for running models) and the model.
- 4Wait while the agent picks the right version for your computer's memory.
- 5Check that the model runs and answers a test question.
- 6Give it your first real task and compare the speed yourself.
Text for your agent
Copy this and send it to your agent — Claude Code, Codex, any of them:
Set up the Swift-Qwen3.8-27B model locally via Ollama: https://huggingface.co/ukisai/Swift-Qwen3.8-27B-GGUF — pick the right quantized variant for my machine, install it, and show me how to use it.
You need a Mac with 32 GB of memory or a GPU with 24 GB of VRAM — the 4-bit version takes about 16 GB.
Open on Hugging Face▌ More finds