README.md

metadata

base_model: Qwen/Qwen3-8B
license: other
license_name: qwen-research
license_link: https://huggingface.co/Qwen/Qwen3-8B/blob/main/LICENSE
model_creator: Qwen
model_name: Qwen3-8B
quantized_by: Second State Inc.

---

Qwen3-8B-GGUF

Original Model

Qwen/Qwen3-8B

Run with LlamaEdge

• LlamaEdge version:

  • Thinking: v0.17.0 and above
  • No Thinking: v0.18.2

• Prompt template

  • Prompt type: chatml (for thinking)

  • Prompt string

    <|im_start|>system
    {system_message}<|im_end|>
    <|im_start|>user
    {prompt}<|im_end|>
    <|im_start|>assistant
    

  • Prompt type: qwen3-no-think (for no thinking)

  • Prompt string

    <|im_start|>system
    {system_message}<|im_end|>
    <|im_start|>user
    {user_message_1}<|im_end|>
    <|im_start|>assistant
    {assistant_message_1}<|im_end|>
    <|im_start|>user
    {user_message_2}<|im_end|>
    <|im_start|>assistant
    <think>
    
    </think>
    

• Context size: 128000

• Run as LlamaEdge service

  wasmedge --dir .:. --nn-preload default:GGML:AUTO:Qwen3-8B-Q5_K_M.gguf \
    llama-api-server.wasm \
    --model-name Qwen3-8B \
    --prompt-template chatml \
    --ctx-size 128000
  

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
Qwen3-8B-Q2_K.gguf	Q2_K	2	3.28 GB	smallest, significant quality loss - not recommended for most purposes
Qwen3-8B-Q3_K_L.gguf	Q3_K_L	3	4.43 GB	small, substantial quality loss
Qwen3-8B-Q3_K_M.gguf	Q3_K_M	3	4.12 GB	very small, high quality loss
Qwen3-8B-Q3_K_S.gguf	Q3_K_S	3	3.77 GB	very small, high quality loss
Qwen3-8B-Q4_0.gguf	Q4_0	4	4.77 GB	legacy; small, very high quality loss - prefer using Q3_K_M
Qwen3-8B-Q4_K_M.gguf	Q4_K_M	4	5.03 GB	medium, balanced quality - recommended
Qwen3-8B-Q4_K_S.gguf	Q4_K_S	4	4.80 GB	small, greater quality loss
Qwen3-8B-Q5_0.gguf	Q5_0	5	5.72 GB	legacy; medium, balanced quality - prefer using Q4_K_M
Qwen3-8B-Q5_K_M.gguf	Q5_K_M	5	5.85 GB	large, very low quality loss - recommended
Qwen3-8B-Q5_K_S.gguf	Q5_K_S	5	5.72 GB	large, low quality loss - recommended
Qwen3-8B-Q6_K.gguf	Q6_K	6	6.73 GB	very large, extremely low quality loss
Qwen3-8B-Q8_0.gguf	Q8_0	8	8.71 GB	very large, extremely low quality loss - not recommended
Qwen3-8B-f16.gguf	f16	16	16.4 GB

Quantized with llama.cpp b5097