Instructions to use ArxhAI/Arxh-2.29-12B-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ArxhAI/Arxh-2.29-12B-it with Transformers:

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

pipe = pipeline("text-generation", model="ArxhAI/Arxh-2.29-12B-it")
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, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("ArxhAI/Arxh-2.29-12B-it")
model = AutoModelForImageTextToText.from_pretrained("ArxhAI/Arxh-2.29-12B-it")
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 ArxhAI/Arxh-2.29-12B-it with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ArxhAI/Arxh-2.29-12B-it"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ArxhAI/Arxh-2.29-12B-it",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ArxhAI/Arxh-2.29-12B-it

SGLang

How to use ArxhAI/Arxh-2.29-12B-it 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 "ArxhAI/Arxh-2.29-12B-it" \
    --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": "ArxhAI/Arxh-2.29-12B-it",
		"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 "ArxhAI/Arxh-2.29-12B-it" \
        --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": "ArxhAI/Arxh-2.29-12B-it",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ArxhAI/Arxh-2.29-12B-it with Docker Model Runner:
```
docker model run hf.co/ArxhAI/Arxh-2.29-12B-it
```

Arxh 2.29 12B

An open-source foundation model for reasoning, coding, and intelligent workflows.

License: Apache 2.0 | Model Family: Arxh | Type: Open-source Large Language Model

Arxh 2.29 14B is a high-performance open-source language model designed for advanced reasoning, technical assistance, code generation, and real-world problem solving. Built as a core model for the Arxh ecosystem, it aims to combine strong capability, clean behavior, and practical deployment efficiency in a form that is accessible to developers, researchers, and the wider open-source community.

Arxh 2.29 14B is engineered to deliver reliable performance across a broad range of language tasks, including structured reasoning, conversational assistance, summarization, rewriting, code understanding, and tool-oriented workflows. The model is intended to serve as a versatile foundation for building intelligent systems that are both capable and adaptable.

Rather than focusing only on fluent text generation, Arxh is designed with an emphasis on useful behavior: following instructions carefully, maintaining clarity in responses, supporting analytical thinking, and integrating naturally into agentic and retrieval-augmented systems.

Design Goals

Arxh 2.29 14B is built around a few core principles:

Strong instruction following
Clear, structured, and professional responses
Reliable reasoning and analytical performance
Practical coding and technical support
Compatibility with tool-using and retrieval-based systems
Flexibility for fine-tuning, experimentation, and deployment

Core Capabilities

Instruction following
Reasoning and problem solving
Code generation and debugging
Summarization and rewriting
Multiturn conversation
Tool use and agent workflows
Retrieval-augmented generation support
Domain-specific adaptation through fine-tuning

Intended Usage

Arxh 2.29 14B is designed for:

Open-source assistant development
Local and cloud deployment
Research and experimentation
Fine-tuning for specialized domains
Developer tools and productivity applications
Agentic AI systems
Knowledge-assisted workflows

Project Vision

Arxh is intended to grow into a family of open models focused on practical intelligence, developer productivity, and accessible AI innovation. The long-term vision is to create a reliable open-source ecosystem where users can build assistants, coding tools, research systems, and autonomous workflows on top of a strong foundation model.

The Arxh project aims to balance openness, capability, and usability so that the model remains useful not only in research environments, but also in real-world applications.

Notes

This repository is part of the broader Arxh model family and may evolve over time through post-training, instruction refinement, domain adaptation, and community-driven improvement.

Downloads last month: 76

Safetensors

Model size

12B params

Tensor type

BF16