---
license: apache-2.0
base_model: Qwen/Qwen3-VL-8B-Instruct
base_model_relation: quantized
tags:
  - gguf
  - vision
  - qwen3-vl
  - llama-cpp
---

# Qwen3-VL-8B-Instruct (GGUF, Q4_K_M)

Mirror of [unsloth/Qwen3-VL-8B-Instruct-GGUF](https://huggingface.co/unsloth/Qwen3-VL-8B-Instruct-GGUF) for use by [local-vision-mcp](https://github.com/froggeric/llm/tree/main/mcp/local-vision-mcp).

## Files

- `Qwen3-VL-8B-Instruct-Q4_K_M.gguf` — main model, 4-bit K-quant (medium), 4.7 GB
- `mmproj-F16.gguf` — vision projector, float16, 1.1 GB

## SHA256

```
Qwen3-VL-8B-Instruct-Q4_K_M.gguf  108e7ff92b78eefd3db4741885104acba514255c11b617d3c7b197a5f46efe89
mmproj-F16.gguf                   d406d03ebabefdef86a2c86bf0c1b65f9e046f7a81c218f25de4931b46a07fc4
```

## License

Apache-2.0 (same as upstream Qwen3-VL-8B-Instruct).