Instructions to use DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking with Transformers:

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

pipe = pipeline("image-text-to-text", model="DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking")
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("DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking")
model = AutoModelForMultimodalLM.from_pretrained("DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking", 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 DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking

SGLang

How to use DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking 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 "DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking" \
    --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": "DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking" \
        --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": "DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Unsloth Studio

How to use DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking",
    max_seq_length=2048,
)

Docker Model Runner
How to use DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking with Docker Model Runner:
```
docker model run hf.co/DavidAU/Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking
```

Gratitude and a question

by Lintrarius - opened May 4

Discussion

Lintrarius

May 4

This model is absolutely superb. Thank you for your work!

I’ve tried all the Qwen 3.5 and 3.6 versions trained on Claude data, but this particular version is the most impressive. It’s the best fit for everyday interaction and creative work. This makes perfect sense: Qwen 3.5, like Opus 4.5, is incomparably “warmer” than 3.6 and 4.6. It’s a perfect combination.

If you ever have the opportunity to release a Qwen 3.5 + Claude 4.5 Opus High Reasoning Thinking model with the default 27B parameters, it would be an invaluable gift for everyone who admires your work but can’t afford to run a 40B model. I’m sure there are many such people. I’d be thrilled to have this model in a lightweight 27B version.

DavidAU

Owner May 5

Thank you ! ;

RE : 27B ;
Part of what makes this (40B) version is the primary training of 27B "base" on Deckard, then expansion, then training again with Claude dataset.
There is a lot going on here aside from these steps - it is the sum of all the steps and how their interact.
One step is the raw creative power, the second extends this creative power and the Claude step unifies and connects (in part) this power together.

The closest 27B is :
https://huggingface.co/DavidAU/Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-Thinking

However; because this model does not have an expansion / 2nd training step it does not have the power of the 40B.

Lintrarius

May 5

Thank you! It’s incredibly interesting to read about the behind‑the‑scenes details.

Yes, I’ve tried models fine‑tuned on DECKARD. But one of the most significant changes after fine‑tuning on Claude’s data is the shift in the model’s “thinking” pattern. It has become much more concise without losing accuracy, “warmer”, less prone to getting stuck in loops, and overall enables the model to structure a more “lively” yet equally logical final response.

This is the real highlight that sets the Qwen 3.5 Claude 4.5 Opus version apart from the base model or any other version. Thank you for it.

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