Instructions to use yasserrmd/AgenticCoder-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use yasserrmd/AgenticCoder-4B with Transformers:

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

pipe = pipeline("text-generation", model="yasserrmd/AgenticCoder-4B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("yasserrmd/AgenticCoder-4B")
model = AutoModelForCausalLM.from_pretrained("yasserrmd/AgenticCoder-4B", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use yasserrmd/AgenticCoder-4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "yasserrmd/AgenticCoder-4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "yasserrmd/AgenticCoder-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/yasserrmd/AgenticCoder-4B

SGLang

How to use yasserrmd/AgenticCoder-4B 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 "yasserrmd/AgenticCoder-4B" \
    --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": "yasserrmd/AgenticCoder-4B",
		"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 "yasserrmd/AgenticCoder-4B" \
        --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": "yasserrmd/AgenticCoder-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use yasserrmd/AgenticCoder-4B with Docker Model Runner:
```
docker model run hf.co/yasserrmd/AgenticCoder-4B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

🧠 AgenticCoder‑4B

AgenticCoder‑4B is a compact 4B parameter language model designed for autonomous agent workflows and intelligent code reasoning. It merges the planning and tool-use strengths of Jan-nano with the coding and logic capabilities of Qwen3‑4B‑Code‑Reasoning, creating a balanced model ideal for real-world assistant scenarios, research agents, and smart development tools.

✨ Key Features

🔁 Agentic Planning & MCP Alignment
Trained on datasets and architectures optimized for multi-step reasoning, task decomposition, and memory–contextual workflows.
💻 Code Understanding & Reasoning
Strong capabilities in Python code generation, script explanation, optimization, and multi-turn task development.
🧰 Tool Use Simulation
Handles realistic tool interaction prompts such as CSV analysis, OCR, and file parsing in code.
📦 Compact & Efficient (4B)
Lightweight enough for cost-efficient deployment, edge device integration, and fine-tuning.

🛠️ Merge Details

Merge Method: SLERP (t = 0.4)
Base Model: Menlo/Jan-nano
Merged With: ertghiu256/qwen3-4b-code-reasoning
Precision: float16
Tokenizer Source: Menlo/Jan-nano

📎 Example Use Cases

✅ "Design a 3-week beginner Python curriculum including AI tools."
✅ "Write a Python function to recursively scan JSON for a key, without using recursion."
✅ "Read a folder of images and extract text using OCR, save to files."
✅ "Summarize trends in a sales CSV and visualize monthly performance."

📁 License & Use

This model is provided for research and development use under the terms of the base models’ respective licenses. Please ensure compliance before commercial usage.

🧬 Citation

If you use this model, consider citing it as:

@misc{agenticcoder4b2025,
  title={AgenticCoder-4B: A Compact Agent + Code Reasoning Model},
  author={Yasser, M.},
  year={2025},
  url={https://huggingface.co/your-username/AgenticCoder-4B}
}

🤝 Acknowledgements

Menlo/Jan-nano by Menlo Systems
Qwen3‑4B‑Code‑Reasoning by ertghiu256
MergeKit, SLERP, Hugging Face

Downloads last month: 145

Safetensors

Model size

4B params

Tensor type

F16

Model tree for yasserrmd/AgenticCoder-4B

Menlo/Jan-nano

ertghiu256/qwen3-4b-code-reasoning

Merge model

this model

Quantizations

3 models