Instructions to use Qwen/Qwen3-4B-Instruct-2507 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Qwen/Qwen3-4B-Instruct-2507 with Transformers:

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

pipe = pipeline("text-generation", model="Qwen/Qwen3-4B-Instruct-2507")
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("Qwen/Qwen3-4B-Instruct-2507")
model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B-Instruct-2507", 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]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Qwen/Qwen3-4B-Instruct-2507 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Qwen/Qwen3-4B-Instruct-2507"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Qwen/Qwen3-4B-Instruct-2507",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Qwen/Qwen3-4B-Instruct-2507

SGLang

How to use Qwen/Qwen3-4B-Instruct-2507 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 "Qwen/Qwen3-4B-Instruct-2507" \
    --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": "Qwen/Qwen3-4B-Instruct-2507",
		"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 "Qwen/Qwen3-4B-Instruct-2507" \
        --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": "Qwen/Qwen3-4B-Instruct-2507",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Qwen/Qwen3-4B-Instruct-2507 with Docker Model Runner:
```
docker model run hf.co/Qwen/Qwen3-4B-Instruct-2507
```

<tool_call> generated even with no tools or asked for

#12

by dipta007 - opened Sep 28, 2025

Discussion

dipta007

Sep 28, 2025

I used the given script for inference with the following prompt. The output always starts with <tool_call> rather than following the specific instruction. And also there is tools are provided, just used the transformer script form the model card.

prompt = """\
You are an expert mathematics assistant. A conversation takes place between the user and the assistant. The user asks a question, and the assistant solves the mathematical problem. The assistant first thinks through the reasoning process in their mind and then provides the answer to the user. The reasoning process and the answer are enclosed within <think> </think> and <answer> </answer> tags respectively.

Your answer must be in the following format:
<think>
Reasoning process here. Reasoning must be in Bengali. Please reason step by step. It can be long and detailed.
</think>
<answer>
Answer here in a single number, must be in Bengali.
</answer>.

Question: Amelia has to do laundry. She has a load of white clothes that will take 72 minutes in the washing machine and 50 minutes in the dryer. She has a load of dark clothes that will take 58 minutes in the washing machine and some time in the dryer. She has a load of colored clothes that will take 45 minutes in the washing machine and 54 minutes in the dryer. Washing and drying all three loads takes her 344 minutes in total. How long does the dark clothes load take in the dryer?
"""

Caydence

Nov 12, 2025

same

Caydence

Nov 12, 2025

same question

txwang98

Nov 17, 2025

same question

Caydence

Nov 17, 2025

Friends, I found a solution. As long as I don't use the 'think' tag, everything works fine. Using 'thinking', 'reasoning', or any tag other than 'think' works completely normally.

txwang98

Nov 17, 2025

Thanks@Caydence. Consider giving it a try.

txwang98

Nov 17, 2025

Friends, I found a solution. As long as I don't use the 'think' tag, everything works fine. Using 'thinking', 'reasoning', or any tag other than 'think' works completely normally.

It works well. Thanks again!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment