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

Libraries

How to use CloudGoat/Orythos-9B with Transformers:

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

pipe = pipeline("image-text-to-text", model="CloudGoat/Orythos-9B")
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("CloudGoat/Orythos-9B")
model = AutoModelForMultimodalLM.from_pretrained("CloudGoat/Orythos-9B", 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

llama.cpp

How to use CloudGoat/Orythos-9B with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf CloudGoat/Orythos-9B:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf CloudGoat/Orythos-9B:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf CloudGoat/Orythos-9B:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf CloudGoat/Orythos-9B:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf CloudGoat/Orythos-9B:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf CloudGoat/Orythos-9B:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf CloudGoat/Orythos-9B:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf CloudGoat/Orythos-9B:Q4_K_M

Use Docker

docker model run hf.co/CloudGoat/Orythos-9B:Q4_K_M

LM Studio
Jan

vLLM

How to use CloudGoat/Orythos-9B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CloudGoat/Orythos-9B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CloudGoat/Orythos-9B",
		"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/CloudGoat/Orythos-9B:Q4_K_M

SGLang

How to use CloudGoat/Orythos-9B 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 "CloudGoat/Orythos-9B" \
    --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": "CloudGoat/Orythos-9B",
		"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 "CloudGoat/Orythos-9B" \
        --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": "CloudGoat/Orythos-9B",
		"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"
						}
					}
				]
			}
		]
	}'

Ollama
How to use CloudGoat/Orythos-9B with Ollama:
```
ollama run hf.co/CloudGoat/Orythos-9B:Q4_K_M
```
Unsloth Desktop

How to use CloudGoat/Orythos-9B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf CloudGoat/Orythos-9B:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "CloudGoat/Orythos-9B:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use CloudGoat/Orythos-9B with Docker Model Runner:
```
docker model run hf.co/CloudGoat/Orythos-9B:Q4_K_M
```

Lemonade

How to use CloudGoat/Orythos-9B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull CloudGoat/Orythos-9B:Q4_K_M

Run and chat with the model

lemonade run user.Orythos-9B-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use CloudGoat/Orythos-9B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf CloudGoat/Orythos-9B:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default CloudGoat/Orythos-9B:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use CloudGoat/Orythos-9B with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf CloudGoat/Orythos-9B:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "CloudGoat/Orythos-9B:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Orythos-9B

While Qwythos is a model specialized in front-end design and reasoning capabilities, it lacks coding and agent-like functionality; therefore, the goal is to port Ornith's agent-like behavior to it.

Merge Details

Merge Method

It was merged using DARE-TIES based on Qwen/Qwen3.5-9B. The merge parameters and method were selected using CloudGoat/RoMM's router.py. The logs from router.py are as follows:

=======================================================================================================================================
[16:37:06] [INFO]  RoMM Statistical Routing Summary [Stable Rank (D_eff)] (Z_slerp: +2.5σ, Z_conflict: -2.0σ, Mean ||v||: 8.0807)
[16:37:06] [INFO] =======================================================================================================================================
[16:37:06] [INFO] Layer  | Cosine      | D_eff    | Z-score   | Category            | Method     | ||dA||  | ||dB||  | Calibrated Parameters
[16:37:06] [INFO] ---------------------------------------------------------------------------------------------------------------------------------------
[16:37:06] [INFO] L0     |  +0.000859 | 1779.4   |   +0.04σ | Orthogonal/Moderate | DARE_TIES  | 8.5413  | 1.5908  | w=[A:0.261, B:0.739] d=0.200
[16:37:06] [INFO] L1     |  +0.001155 | 1637.9   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 8.7600  | 1.2983  | w=[A:0.246, B:0.754] d=0.200
[16:37:06] [INFO] L2     |  +0.001124 | 1729.0   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 9.0479  | 1.2902  | w=[A:0.241, B:0.759] d=0.200
[16:37:06] [INFO] L3     |  +0.001131 | 1310.2   |   +0.04σ | Orthogonal/Moderate | DARE_TIES  | 9.2043  | 1.4826  | w=[A:0.245, B:0.755] d=0.200
[16:37:06] [INFO] L4     |  +0.001321 | 1562.7   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 9.9421  | 1.2857  | w=[A:0.227, B:0.773] d=0.200
[16:37:06] [INFO] L5     |  +0.001187 | 1677.0   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 10.3982 | 1.3783  | w=[A:0.223, B:0.777] d=0.200
[16:37:06] [INFO] L6     |  +0.000930 | 1629.4   |   +0.04σ | Orthogonal/Moderate | DARE_TIES  | 10.8374 | 1.7191  | w=[A:0.229, B:0.771] d=0.200
[16:37:06] [INFO] L7     |  +0.001257 | 2028.2   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 11.0917 | 3.5998  | w=[A:0.297, B:0.703] d=0.200
[16:37:06] [INFO] L8     |  +0.001245 | 2318.5   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 12.0419 | 3.6030  | w=[A:0.283, B:0.717] d=0.200
[16:37:06] [INFO] L9     |  +0.001184 | 2306.0   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 12.4915 | 3.7002  | w=[A:0.279, B:0.721] d=0.200
[16:37:06] [INFO] L10    |  +0.001270 | 1974.7   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 12.6918 | 3.7674  | w=[A:0.279, B:0.721] d=0.200
[16:37:06] [INFO] L11    |  +0.001302 | 2190.4   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 12.7971 | 3.8055  | w=[A:0.279, B:0.721] d=0.200
[16:37:06] [INFO] L12    |  +0.001218 | 1899.5   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 13.6662 | 3.8524  | w=[A:0.269, B:0.731] d=0.200
[16:37:06] [INFO] L13    |  +0.001360 | 1874.0   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 14.0397 | 3.8960  | w=[A:0.266, B:0.734] d=0.200
[16:37:06] [INFO] L14    |  +0.001320 | 1620.0   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 13.9480 | 3.8657  | w=[A:0.266, B:0.734] d=0.200
[16:37:06] [INFO] L15    |  +0.001323 | 1661.6   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 13.8683 | 3.9140  | w=[A:0.268, B:0.732] d=0.200
[16:37:06] [INFO] L16    |  +0.001287 | 1296.4   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 14.7141 | 3.8759  | w=[A:0.257, B:0.743] d=0.200
[16:37:06] [INFO] L17    |  +0.001521 | 1421.4   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 15.0858 | 3.8054  | w=[A:0.250, B:0.750] d=0.200
[16:37:06] [INFO] L18    |  +0.001550 | 1518.2   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 14.7944 | 3.6716  | w=[A:0.249, B:0.751] d=0.200
[16:37:06] [INFO] L19    |  +0.001532 | 1902.5   |   +0.07σ | Orthogonal/Moderate | DARE_TIES  | 14.4013 | 3.8580  | w=[A:0.260, B:0.740] d=0.200
[16:37:06] [INFO] L20    |  +0.001487 | 1690.8   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 14.8436 | 3.7270  | w=[A:0.251, B:0.749] d=0.200
[16:37:06] [INFO] L21    |  +0.001540 | 1562.5   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 14.9207 | 3.7038  | w=[A:0.249, B:0.751] d=0.200
[16:37:06] [INFO] L22    |  +0.001554 | 1730.2   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 14.7455 | 3.7406  | w=[A:0.252, B:0.748] d=0.200
[16:37:06] [INFO] L23    |  +0.001576 | 1751.1   |   +0.07σ | Orthogonal/Moderate | DARE_TIES  | 14.2006 | 3.8843  | w=[A:0.263, B:0.737] d=0.200
[16:37:06] [INFO] L24    |  +0.001638 | 1823.6   |   +0.07σ | Orthogonal/Moderate | DARE_TIES  | 14.2273 | 3.7918  | w=[A:0.260, B:0.740] d=0.200
[16:37:06] [INFO] L25    |  +0.001520 | 1844.2   |   +0.07σ | Orthogonal/Moderate | DARE_TIES  | 14.1245 | 3.7967  | w=[A:0.261, B:0.739] d=0.200
[16:37:06] [INFO] L26    |  +0.001268 | 1750.7   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 13.9964 | 3.7873  | w=[A:0.263, B:0.737] d=0.200
[16:37:06] [INFO] L27    |  +0.001319 | 1692.1   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 14.1992 | 4.1041  | w=[A:0.270, B:0.730] d=0.200
[16:37:06] [INFO] L28    |  +0.001437 | 1678.2   |   +0.06σ | Orthogonal/Moderate | DARE_TIES  | 14.1633 | 3.7930  | w=[A:0.261, B:0.739] d=0.200
[16:37:06] [INFO] L29    |  +0.001411 | 1195.6   |   +0.05σ | Orthogonal/Moderate | DARE_TIES  | 13.8456 | 3.8167  | w=[A:0.265, B:0.735] d=0.200
[16:37:06] [INFO] L30    |  +0.001404 | 793.4    |   +0.04σ | Orthogonal/Moderate | DARE_TIES  | 13.9985 | 3.8576  | w=[A:0.265, B:0.735] d=0.200
[16:37:06] [INFO] L31    |  +0.001631 | 563.6    |   +0.04σ | Orthogonal/Moderate | DARE_TIES  | 12.4993 | 3.7773  | w=[A:0.282, B:0.718] d=0.200
[16:37:06] [INFO] =======================================================================================================================================

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

merge_method: dare_ties
base_model: Qwen/Qwen3.5-9B
dtype: bfloat16
parameters:
  density: 0.2
models:
- model: ornith-ai/Ornith-1.5-9B
  parameters:
    weight:
    - filter: layers.0.
      value: 0.2612
    - filter: layers.1.
      value: 0.2464
    - filter: layers.2.
      value: 0.2412
    - filter: layers.3.
      value: 0.2453
    - filter: layers.4.
      value: 0.2268
    - filter: layers.5.
      value: 0.2231
    - filter: layers.6.
      value: 0.2289
    - filter: layers.7.
      value: 0.2971
    - filter: layers.8.
      value: 0.2826
    - filter: layers.9.
      value: 0.2795
    - filter: layers.10.
      value: 0.279
    - filter: layers.11.
      value: 0.2789
    - filter: layers.12.
      value: 0.2689
    - filter: layers.13.
      value: 0.2656
    - filter: layers.14.
      value: 0.2658
    - filter: layers.15.
      value: 0.2684
    - filter: layers.16.
      value: 0.2569
    - filter: layers.17.
      value: 0.2504
    - filter: layers.18.
      value: 0.2493
    - filter: layers.19.
      value: 0.26
    - filter: layers.20.
      value: 0.2506
    - filter: layers.21.
      value: 0.249
    - filter: layers.22.
      value: 0.2521
    - filter: layers.23.
      value: 0.2633
    - filter: layers.24.
      value: 0.2599
    - filter: layers.25.
      value: 0.2613
    - filter: layers.26.
      value: 0.2626
    - filter: layers.27.
      value: 0.2704
    - filter: layers.28.
      value: 0.2607
    - filter: layers.29.
      value: 0.2654
    - filter: layers.30.
      value: 0.2649
    - filter: layers.31.
      value: 0.2821
    - value: 0.2599
- model: empero-ai/Qwythos-9B-Claude-Mythos-5-1M
  parameters:
    weight:
    - filter: layers.0.
      value: 0.7388
    - filter: layers.1.
      value: 0.7536
    - filter: layers.2.
      value: 0.7588
    - filter: layers.3.
      value: 0.7547
    - filter: layers.4.
      value: 0.7732
    - filter: layers.5.
      value: 0.7769
    - filter: layers.6.
      value: 0.7711
    - filter: layers.7.
      value: 0.7029
    - filter: layers.8.
      value: 0.7174
    - filter: layers.9.
      value: 0.7205
    - filter: layers.10.
      value: 0.721
    - filter: layers.11.
      value: 0.7211
    - filter: layers.12.
      value: 0.7311
    - filter: layers.13.
      value: 0.7344
    - filter: layers.14.
      value: 0.7342
    - filter: layers.15.
      value: 0.7316
    - filter: layers.16.
      value: 0.7431
    - filter: layers.17.
      value: 0.7496
    - filter: layers.18.
      value: 0.7507
    - filter: layers.19.
      value: 0.74
    - filter: layers.20.
      value: 0.7494
    - filter: layers.21.
      value: 0.751
    - filter: layers.22.
      value: 0.7479
    - filter: layers.23.
      value: 0.7367
    - filter: layers.24.
      value: 0.7401
    - filter: layers.25.
      value: 0.7387
    - filter: layers.26.
      value: 0.7374
    - filter: layers.27.
      value: 0.7296
    - filter: layers.28.
      value: 0.7393
    - filter: layers.29.
      value: 0.7346
    - filter: layers.30.
      value: 0.7351
    - filter: layers.31.
      value: 0.7179
    - value: 0.7401

Downloads last month: -

Safetensors

Model size

9B params

Tensor type

BF16

Model tree for CloudGoat/Orythos-9B

Qwen/Qwen3.5-9B

empero-ai/Qwythos-9B-Claude-Mythos-5-1M

ornith-ai/Ornith-1.5-9B

Merge model

this model

Quantizations

1 model