Instructions to use CosmossG/COSMOS-9B-V1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CosmossG/COSMOS-9B-V1 with Transformers:

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

pipe = pipeline("image-text-to-text", model="CosmossG/COSMOS-9B-V1")
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("CosmossG/COSMOS-9B-V1")
model = AutoModelForMultimodalLM.from_pretrained("CosmossG/COSMOS-9B-V1", 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 CosmossG/COSMOS-9B-V1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CosmossG/COSMOS-9B-V1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CosmossG/COSMOS-9B-V1",
		"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/CosmossG/COSMOS-9B-V1

SGLang

How to use CosmossG/COSMOS-9B-V1 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 "CosmossG/COSMOS-9B-V1" \
    --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": "CosmossG/COSMOS-9B-V1",
		"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 "CosmossG/COSMOS-9B-V1" \
        --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": "CosmossG/COSMOS-9B-V1",
		"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"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use CosmossG/COSMOS-9B-V1 with Docker Model Runner:
```
docker model run hf.co/CosmossG/COSMOS-9B-V1
```

COSMOS-9B-V1 / README.md

CosmossG

Upload folder using huggingface_hub

86acfe1 verified 7 days ago

preview code

Raw

History Blame Contribute Delete

1.31 kB

	---
	language:
	- en
	- pt
	tags:
	- merge
	- dare-ties
	- qwen3.5
	- code
	- reasoning
	- long-context
	base_model:
	- empero-ai/Qwythos-9B-Claude-Mythos-5-1M
	- Tesslate/OmniCoder-9B
	library_name: transformers
	---

	# COSMOS 9B V1 (Base Merge)

	COSMOS 9B V1 is a merged base model combining the advanced reasoning and massive context window of Qwythos-9B with the agentic coding capabilities of OmniCoder-9B.

	## Merge Details

	### Merge Method
	This model was merged using a custom DARE-TIES implementation (50% density, 50% weight) to preserve the distinct specializations of both parent models.

	### Models Merged
	- [empero-ai/Qwythos-9B-Claude-Mythos-5-1M]: A full-parameter reasoning model post-trained on Claude Mythos traces with a 1,048,576-token context window (YaRN).
	- [Tesslate/OmniCoder-9B]: A coding agent model fine-tuned on 425,000+ curated agentic coding trajectories from Claude Opus 4.6.

	### Configuration
	The merge was anchored to the Qwythos-9B base to strictly preserve the YaRN 1M context window, native function calling, and the Qwen3.5 hybrid architecture (Gated Delta Networks).

	## Intended Use
	This is a base merge intended for further fine-tuning, specifically for educational purposes, programming instruction (C/C++), and Portuguese language specialization.