Instructions to use Qwen/Qwen3-Coder-30B-A3B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Qwen/Qwen3-Coder-30B-A3B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="Qwen/Qwen3-Coder-30B-A3B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-Coder-30B-A3B-Instruct")
model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-Coder-30B-A3B-Instruct")
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
Local Apps

vLLM

How to use Qwen/Qwen3-Coder-30B-A3B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Qwen/Qwen3-Coder-30B-A3B-Instruct"
# 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-Coder-30B-A3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Qwen/Qwen3-Coder-30B-A3B-Instruct

SGLang

How to use Qwen/Qwen3-Coder-30B-A3B-Instruct 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-Coder-30B-A3B-Instruct" \
    --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-Coder-30B-A3B-Instruct",
		"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-Coder-30B-A3B-Instruct" \
        --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-Coder-30B-A3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Qwen/Qwen3-Coder-30B-A3B-Instruct with Docker Model Runner:
```
docker model run hf.co/Qwen/Qwen3-Coder-30B-A3B-Instruct
```

Update Jinja template for Jinja version compatibility

#13

by rshcao - opened Aug 7, 2025

base: refs/heads/main

←

from: refs/pr/13

Discussion Files changed

+313

-443

This PR is in draft mode

rshcao

Qwen org Aug 7, 2025

No description provided.

JamesMowery

Aug 7, 2025

•

edited Aug 7, 2025

Any way to get a raw copy of this template? I'm hoping it might help resolve issues I've had with tool calling in tools like RooCode. (Unless it's the same one from Unsloth: it still doesn't work well in RooCode with Llama.cpp.)

keick

Aug 7, 2025

Having the same issues. Tool calling is all wrong and hasn't worked once. Running Ollama 0.11.3, RooCode 3.25.9. And the tool calling is missing the outer wrappers for tool calling.

Example Output:

Let me read one of these files to understand its structure and then document it according to the template. I'll start with the first one.
<read_file path="Databases/Queries/avgFltHrs.md"/>

Whereas devstral produces the correct tool use:

Let me read one of these files to understand its structure and then document it according to the template. I'll start with the first one.
<read_file>
<args>
<file>
<path="Databases/Queries/avgFltHrs.md"/>
</file>
</args>
</read_file>

rshcao

Qwen org Aug 7, 2025

The XML-format raw output for Qwen should work like this: (assistant output for the case you provided above)

<tool_call>
<function=read_file>
<parameter=path>
Databases/Queries/avgFltHrs.md
</parameter>
</function>
</tool_call>

Notice that:

LLM does not need to predict the double quotes for path "Databases/Queries/avgFltHrs.md". This can prevent the notorious problem of escaping characters in multi-line code.
We constrain XML tags only to the special format <function=xxx></function> and <parameter=xxx></parameter>.
That is, we do not use the function name or parameter name itself as XML tags (like the case you provided <read_file> and <file>) to prevent the potential risk incurred by arbitrary tags definition. Arbitrary XML tags will easily cause nested parsing conflicts or difficulty when the parameter value itself is XML code.

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

Tap or paste here to upload images

Publish this branch

This branch is in draft mode, publish it to be able to merge.

· Sign up or log in to comment