---
tags:
  - gguf
  - llama.cpp
  - quantized
  - qwen
  - text-generation
  - local-llm
---
# Qwen3.5-0.8B GGUF Quantizations

Converted from: Qwen/Qwen3.5-0.8B-Base

## Quantizations

- Q2_K → smallest, fastest
- Q3_K_M → balanced
- Q4_K_M → recommended
- Q5_K_M → highest quality

Recommended: Q4_K_M (best balance of speed and quality)

Q2_K → ~200MB

Q4_K_M → ~500MB

Q5_K_M → ~650MB

## Tested on
- LM Studio ✔
- llama.cpp ✔

## Notes
- Converted using llama.cpp
- No LoRA / base model only

---
license: apache 2.0
---