---
license: apache-2.0
base_model:
  - Qwen/Qwen3.6-35B-A3B-FP8
pipeline_tag: text-generation
library_name: transformers
tags:
  - coding
  - agent
  - vision
  - gguf
  - qwen3
  - moe
---


# Lumix-35B

**Lumix-35B** is a vision-capable coding and reasoning model from **Blackfrost**, distributed as GGUF quantizations for [llama.cpp](https://github.com/ggml-org/llama.cpp) and [Ollama](https://ollama.com). It is a **fine-tune of [Qwen3.6-35B-A3B](https://huggingface.co/Qwen/Qwen3.6-35B-A3B-FP8)** (Mixture-of-Experts), tuned to drive the [Shadow](https://github.com/Blackfrost-AI/Shadow_CLI) coding agent.

- **Base:** [Qwen/Qwen3.6-35B-A3B-FP8](https://huggingface.co/Qwen/Qwen3.6-35B-A3B-FP8) (GGUF architecture `qwen35moe`)
- **Parameters:** ~35B total (Mixture-of-Experts, a few B active per token)
- **Modalities:** text **+ vision** (image input via the bundled `mmproj`)
- **Context:** long-context, agentic coding — one clean tool call at a time, look → act → verify

## Files

| File | Quant | Notes |
|---|---|---|
| `LUMIX-35B-Q4_K_M.gguf` | Q4_K_M | **recommended** |
| `LUMIX-35B-Q5_K_M.gguf` | Q5_K_M | |
| `LUMIX-35B-Q6_K.gguf` | Q6_K | |
| `LUMIX-35B-Q8_0.gguf` | Q8_0 | near-lossless |
| `LUMIX-35B-f16.gguf` | F16 | full precision |
| `Lumix-35B-mmproj.gguf` | — | vision projector (pair with any quant for image input) |

## Usage

**llama.cpp**
```bash
llama-cli -hf Blackfrost-AI/Lumix-35B:Q4_K_M
# vision: also pass --mmproj Lumix-35B-mmproj.gguf and an --image
```

**Ollama**
```bash
ollama run hf.co/Blackfrost-AI/Lumix-35B:Q4_K_M
```

## License

Released under the **Apache-2.0** license.