Instructions to use Qwen/Qwen-AgentWorld-35B-A3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Qwen/Qwen-AgentWorld-35B-A3B with Transformers:

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

pipe = pipeline("text-generation", model="Qwen/Qwen-AgentWorld-35B-A3B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("Qwen/Qwen-AgentWorld-35B-A3B")
model = AutoModelForMultimodalLM.from_pretrained("Qwen/Qwen-AgentWorld-35B-A3B", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Qwen/Qwen-AgentWorld-35B-A3B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Qwen/Qwen-AgentWorld-35B-A3B"
# 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/Qwen-AgentWorld-35B-A3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Qwen/Qwen-AgentWorld-35B-A3B

SGLang

How to use Qwen/Qwen-AgentWorld-35B-A3B 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/Qwen-AgentWorld-35B-A3B" \
    --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/Qwen-AgentWorld-35B-A3B",
		"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/Qwen-AgentWorld-35B-A3B" \
        --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/Qwen-AgentWorld-35B-A3B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Qwen/Qwen-AgentWorld-35B-A3B with Docker Model Runner:
```
docker model run hf.co/Qwen/Qwen-AgentWorld-35B-A3B
```

Great on HERMES! Noob Question

by GravyLuver - opened Jun 25

Discussion

GravyLuver

Jun 25

why not use the 3.6 35B base A3B model? instead of the 3.5 35B A3B base model? just want to know the reasoning behind it.

sld26t

Jun 26

It was probably a parallel project run by a separate team, started before Qwen 3.6 launched, my only guess.

inikishev

Jun 27

•

edited Jun 27

why not use the 3.6 35B base A3B model? instead of the 3.5 35B A3B base model? just want to know the reasoning behind it.

There are no 3.6 base models, I think that 3.6 is also retrained from 3.5 base or it could be 3.5 with extra training

Wladastic

Jun 28

I highly suspect 3.6 to be slightly overcooked anyway.
3.5 understands when you tell it that the new us president is NOT Biden anymore as developer or system.
But 3.6 does not even believe the system prompt

Tariel

Jul 1

why not use the 3.6 35B base A3B model? instead of the 3.5 35B A3B base model? just want to know the reasoning behind it.

There is no such a thing as "3.6 base model". Qwen 3.6 35B-A3B/27B models are virtually reinforcement trained versions of corresponding 3.5 models, they have exactly the same structure and tokenizers as 3.5.

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