Instructions to use ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast with Transformers:

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

pipe = pipeline("text-generation", model="ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast")
model = AutoModelForCausalLM.from_pretrained("ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast

SGLang

How to use ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast 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 "ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast" \
    --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": "ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast",
		"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 "ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast" \
        --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": "ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast with Docker Model Runner:
```
docker model run hf.co/ArliAI/Qwen3-30B-A3B-ArliAI-RpR-v4-Fast
```

A little observation

by Danioken - opened Jul 21, 2025

Discussion

Danioken

Jul 21, 2025

•

edited Jul 21, 2025

"RpR models does not work well with repetition penalty type of samplers, even more advanced ones such as XTC or DRY."

While using your model, however, I noticed that setting the Repetition Penalty to 1.1 and the Rep Pen Range to just 64 not only eliminates the looping during reasoning but, in my opinion, significantly improves its quality.

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