Instructions to use HuggingFaceTB/SmolLM3-3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HuggingFaceTB/SmolLM3-3B with Transformers:

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

pipe = pipeline("text-generation", model="HuggingFaceTB/SmolLM3-3B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/SmolLM3-3B")
model = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM3-3B", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use HuggingFaceTB/SmolLM3-3B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HuggingFaceTB/SmolLM3-3B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceTB/SmolLM3-3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/HuggingFaceTB/SmolLM3-3B

SGLang

How to use HuggingFaceTB/SmolLM3-3B 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 "HuggingFaceTB/SmolLM3-3B" \
    --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": "HuggingFaceTB/SmolLM3-3B",
		"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 "HuggingFaceTB/SmolLM3-3B" \
        --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": "HuggingFaceTB/SmolLM3-3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use HuggingFaceTB/SmolLM3-3B with Docker Model Runner:
```
docker model run hf.co/HuggingFaceTB/SmolLM3-3B
```

SmolLM3-3B

Commit History

Jul 9, 2025

12:44
Adding notebook.ipynb for showing inference using Colab⁠

sergiopaniego HF Staff

Verified signature
f22368b
11:10
smol markup error (#11)⁠

loubnabnl +1: burtenshaw HF Staff

Verified signature
66fc337
10:05
fix: reassign xml_tools to tools to make it work with vllm (#9)⁠

XciD HF Staff

Verified signature
486a602

Jul 8, 2025

22:00
Update README.md

eliebak HF Staff

Verified signature
b49c771
15:54
fix todo

eliebak HF Staff

Verified signature
90486a3
15:23
Update README.md

eliebak HF Staff

Verified signature
f17cc5c
14:35
Update README.md

m-ric

Verified signature
ce6f3a2
14:31
Update README.md

m-ric

Verified signature
78ed62d
14:31
Update README.md

eliebak HF Staff

Verified signature
ae6225b
14:30
Update README.md (#4)⁠

eliebak HF Staff

Verified signature
2585aa5
14:28
Update chat_template.jinja

m-ric

Verified signature
a169fa4
14:23
Update generation_config.json (#3)⁠

cmpatino +1: Xenova HF Staff

Verified signature
2ea4d77
14:23
Update README.md

nouamanetazi HF Staff

Verified signature
ddd1ea1
14:03
Update README.md

edbeeching HF Staff

Verified signature
1ca9692
14:01
Update README.md

edbeeching HF Staff

Verified signature
bbd8497
13:49
add how to use (#2)⁠

loubnabnl HF Staff

Verified signature
0831828
13:45
Update README.md

eliebak HF Staff

Verified signature
9d66358
13:43
Update README.md

eliebak HF Staff

Verified signature
2593281
13:39
Add vllm and slang commands

cmpatino HF Staff

aa9d5a1
13:27
update evaluation section and remove extra space (#1)⁠

cmpatino +1: loubnabnl HF Staff

Verified signature
d4ed61a
13:22
Update README.md

eliebak HF Staff

Verified signature
320cd3e
13:15
Escape <code> tag in the README

cmpatino HF Staff

c48d82f
13:13
Add examples for using the model

cmpatino HF Staff

1b310b4
12:59
Update README.md

m-ric

Verified signature
6e4efe2
12:59
Update README.md

m-ric

Verified signature
4d66e40
12:58
Update README.md

m-ric

Verified signature
e67b652
10:15
Update EOS token ID

cmpatino HF Staff

Verified signature
82547f5
10:13
Edit model card

cmpatino HF Staff

Verified signature
d591279
10:12
Upload folder using huggingface_hub

cmpatino HF Staff

Verified signature
451a3a6
10:11
initial commit

cmpatino HF Staff

Verified signature
f5e14ae

Commit History

Jul 9, 2025

Adding notebook.ipynb for showing inference using Colab⁠

smol markup error (#11)⁠

fix: reassign xml_tools to tools to make it work with vllm (#9)⁠

Jul 8, 2025

Update README.md

fix todo

Update README.md

Update README.md

Update README.md

Update README.md

Update README.md (#4)⁠

Update chat_template.jinja

Update generation_config.json (#3)⁠

Update README.md

Update README.md

Update README.md

add how to use (#2)⁠

Update README.md

Update README.md

Add vllm and slang commands

update evaluation section and remove extra space (#1)⁠

Update README.md

Escape <code> tag in the README

Add examples for using the model

Update README.md

Update README.md

Update README.md

Update EOS token ID

Edit model card

Upload folder using huggingface_hub

initial commit