---
base_model: openai/gpt-oss-20b
license: apache-2.0
model_creator: openai
model_name: gpt-oss-20b
quantized_by: Second State Inc.
pipeline_tag: text-generation
library_name: transformers
---

<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->

# gpt-oss-20b-GGUF

## Original Model

[openai/gpt-oss-20b](https://huggingface.co/openai/gpt-oss-20b)

## Run with LlamaEdge

- LlamaEdge version: [v0.25.0](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.25.0) and above ([0.25.1](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.25.1)+ with tool call support)

- Prompt template

  - Prompt type: `gpt-oss`

  - Prompt string

    ```text
    <|start|>system<|message|>
    You are ChatGPT, a large language model trained by OpenAI.
    Knowledge cutoff: 2024-06
    Current date: 2025-08-06
    Reasoning: medium
    # Valid channels: analysis, commentary, final. Channel must be included for every message.
    <|end|>

    <|start|>user<|message|>Hello!<|end|>
    <|start|>assistant<|channel|>final<|message|>Hi there!<|end|>
    <|start|>user<|message|>What's your favorite color?<|end|>
    <|start|>assistant
    ```

- Context size: `128000`

- Run as LlamaEdge service

  ```bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:gpt-oss-20b-MXFP4_MOE.gguf \
    llama-api-server.wasm \
    --model-name gpt-oss-20b \
    --prompt-template gpt-oss \
    --ctx-size 128000
  ```

<!-- ## Quantized GGUF Models

| Name | Quant method | Bits | Size | Use case |
| ---- | ---- | ---- | ---- | ----- |
| [gpt-oss-20b-Q2_K.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q2_K.gguf)     | Q2_K   | 2 | 968 MB| smallest, significant quality loss - not recommended for most purposes |
| [gpt-oss-20b-Q3_K_L.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q3_K_L.gguf) | Q3_K_L | 3 | 971 MB| small, substantial quality loss |
| [gpt-oss-20b-Q3_K_M.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q3_K_M.gguf) | Q3_K_M | 3 | 970 MB| very small, high quality loss |
| [gpt-oss-20b-Q3_K_S.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q3_K_S.gguf) | Q3_K_S | 3 | 968 MB| very small, high quality loss |
| [gpt-oss-20b-Q4_0.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q4_0.gguf)     | Q4_0   | 4 | 969 MB| legacy; small, very high quality loss - prefer using Q3_K_M |
| [gpt-oss-20b-Q4_K_M.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q4_K_M.gguf) | Q4_K_M | 4 | 1.04 GB| medium, balanced quality - recommended |
| [gpt-oss-20b-Q4_K_S.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q4_K_S.gguf) | Q4_K_S | 4 | 1.04 GB| small, greater quality loss |
| [gpt-oss-20b-Q5_0.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q5_0.gguf)     | Q5_0   | 5 | 1.05 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
| [gpt-oss-20b-Q5_K_M.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q5_K_M.gguf) | Q5_K_M | 5 | 1.08 GB| large, very low quality loss - recommended |
| [gpt-oss-20b-Q5_K_S.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q5_K_S.gguf) | Q5_K_S | 5 | 1.08 GB| large, low quality loss - recommended |
| [gpt-oss-20b-Q6_K.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q6_K.gguf)     | Q6_K   | 6 | 1.27 GB| very large, extremely low quality loss |
| [gpt-oss-20b-Q8_0.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-Q8_0.gguf)     | Q8_0   | 8 | 1.27 GB| very large, extremely low quality loss - not recommended |
| [gpt-oss-20b-f16.gguf](https://huggingface.co/second-state/gpt-oss-20b-GGUF/blob/main/gpt-oss-20b-f16.gguf)       | f16    | 16 | 13.8 GB| | -->

*Quantized with llama.cpp b6115*