openai
/

gpt-oss-20b

Text Generation

8-bit precision

Model card Files Files and versions

Instructions to use openai/gpt-oss-20b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use openai/gpt-oss-20b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="openai/gpt-oss-20b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("openai/gpt-oss-20b")
model = AutoModelForCausalLM.from_pretrained("openai/gpt-oss-20b", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
AMD Developer Cloud
Local Apps Settings

How to use openai/gpt-oss-20b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "openai/gpt-oss-20b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openai/gpt-oss-20b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/openai/gpt-oss-20b

How to use openai/gpt-oss-20b with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "openai/gpt-oss-20b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openai/gpt-oss-20b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "openai/gpt-oss-20b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openai/gpt-oss-20b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use openai/gpt-oss-20b with Docker Model Runner:
```
docker model run hf.co/openai/gpt-oss-20b
```

Resources

View closed (39)

off111README.md

#168 opened 8 months ago by

Appreciation for GPT-OSS-20B and Inquiry on Future Vision-Language Models

#167 opened 9 months ago by

Small changes in readme for Vllm serve

#166 opened 9 months ago by

No matching torch==2.9.0.dev20250804+cu128 wheel for Python 3.12 – cannot install vllm==0.10.1+gptoss with official instructions

#165 opened 9 months ago by

is that your page? its a copy

#164 opened 9 months ago by

metal example code doesn't run as written

#163 opened 9 months ago by

Update README.md

#162 opened 9 months ago by

hedi-huggingface

Genuine User Reviews and Questions on Repo openai/gpt-oss-20b

#161 opened 9 months ago by

Tool Calling in Chat Template

#160 opened 10 months ago by

Update README.md

#159 opened 10 months ago by

Miss

#158 opened 10 months ago by deleted

deploying finetuned model on triton using vllm backend

#157 opened 10 months ago by

gpt

#156 opened 10 months ago by

Missing torchvision==0.24.0.dev20250804+cu128 required by vllm+gptoss installation

#155 opened 10 months ago by

Request: DOI

#154 opened 10 months ago by

Reinforcement Learning example

#153 opened 10 months ago by

is the model support TGI?

#152 opened 10 months ago by

Tokenizer vocab size doesn't match model vocab size

#151 opened 10 months ago by

Request: DOI

#150 opened 11 months ago by

Cannot retrieve streamed reasoning output from openai/gpt-oss-20b with vLLM and LangChain

#149 opened 11 months ago by

Database Intelligent Query System Implemented Using the GPT-OSS-20B Model

#148 opened 11 months ago by

Inference engines that actually use OSS to full extent? (responses, harmony)

#147 opened 11 months ago by

Create requirements.txt

#146 opened 11 months ago by

skminhajuddin20

Create app.py

#145 opened 11 months ago by

skminhajuddin20

Update Readme.md add multilingual language tag

#143 opened 11 months ago by

"Karen from HR has called... she wants her LLM back."

#142 opened 11 months ago by

builderofworlds

Speculative Decoders?

#141 opened 11 months ago by

Privacy moment.

#138 opened 12 months ago by

test

#137 opened 12 months ago by

Unable to load gpt-oss-20b on dual L40 (48GB) GPUs with vLLM

#136 opened 12 months ago by

int32 overflow detected for operation add in `_keyed_add` kernel under TRITON_DEBUG=1

#135 opened 12 months ago by

myllm

#134 opened 12 months ago by

zzsm

#132 opened 12 months ago by

1111

#130 opened 12 months ago by

Update README.md

#129 opened 12 months ago by

My friend and I asked the exact same question but why was his approved but mine refused?

#127 opened 12 months ago by

Wrap assistant messages inside "{% generation %}" markers in chat_template.jinja

#126 opened 12 months ago by

Upload medical_diagnosis_manual-1.pdf

#125 opened 12 months ago by

Not able to deploy gpt-oss-20b model in A100s

#124 opened 12 months ago by

20B Parameters vs ChatGPT 4

#123 opened 12 months ago by

68.4 on Aider Polyglot with reasoning_effort: high

#122 opened 12 months ago by

MXFP4 utilization over NVFP4

#121 opened 12 months ago by

gpt-oss-20b Nvdidia A10G error deploy Attention Sinks

#120 opened 12 months ago by

The output of `gpt_oss.generate` is meaningless

#119 opened 12 months ago by

running gpt-oss-20B on Intel Arc iGPU

#117 opened 12 months ago by

Update chat_template.jinja to correctly parse types nested in arrays

#116 opened 12 months ago by

Model removed from openrouter.ai

#115 opened 12 months ago by

Performance gains on low-end hardware

#114 opened 12 months ago by

Is it mandatory to install flash-attention for GPT-OSS?

#112 opened 12 months ago by

Unable to Structured output

#111 opened 12 months ago by