Instructions to use Neura-Tech-AI/Nexa-AI-4x4B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Neura-Tech-AI/Nexa-AI-4x4B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="Neura-Tech-AI/Nexa-AI-4x4B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Neura-Tech-AI/Nexa-AI-4x4B-Instruct")
model = AutoModelForCausalLM.from_pretrained("Neura-Tech-AI/Nexa-AI-4x4B-Instruct", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Neura-Tech-AI/Nexa-AI-4x4B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Neura-Tech-AI/Nexa-AI-4x4B-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Neura-Tech-AI/Nexa-AI-4x4B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Neura-Tech-AI/Nexa-AI-4x4B-Instruct

SGLang

How to use Neura-Tech-AI/Nexa-AI-4x4B-Instruct 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 "Neura-Tech-AI/Nexa-AI-4x4B-Instruct" \
    --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": "Neura-Tech-AI/Nexa-AI-4x4B-Instruct",
		"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 "Neura-Tech-AI/Nexa-AI-4x4B-Instruct" \
        --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": "Neura-Tech-AI/Nexa-AI-4x4B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Neura-Tech-AI/Nexa-AI-4x4B-Instruct with Docker Model Runner:
```
docker model run hf.co/Neura-Tech-AI/Nexa-AI-4x4B-Instruct
```

Nexa-AI-4x4B-Instruct / NOTICE

Neura-Tech-AI

Update NOTICE

06117ea verified 18 days ago

Raw

History Blame Contribute Delete

1.53 kB

	Nexa-AI-4x4B-Instruct

	Copyright (c) 2026 Neura Tech AI & Lumina AI

	Nexa-AI-4x4B-Instruct is a collaborative open-source large language model
	developed by Neura Tech AI and Lumina AI.

	This model incorporates and builds upon components derived from:

	- Neura-Tech-AI/Nexa-AI-4B-Instruct
	- Qwen/Qwen3-4B-Instruct-2507
	- Qwen/Qwen3-4B-Thinking-2507

	Copyright (c) 2024 Alibaba Cloud and the Qwen Team.

	The original Qwen models are licensed under the Apache License, Version 2.0.
	A copy of the Apache License is included in the LICENSE file.

	Modifications, integrations, and additions made by Neura Tech AI & Lumina AI include:

	- Mixture of Experts (MoE) architecture development
	- Expert routing and sparse computation design
	- Expert integration and model composition
	- Instruction tuning and alignment improvements
	- Reasoning capability enhancements
	- Identity and behavioral customization
	- Chat template customization
	- Dataset curation and quality improvements
	- Multilingual optimization
	- Tokenizer and configuration adjustments
	- Inference optimization
	- Model packaging, documentation, and distribution
	- Branding and project identity

	These modifications are Copyright (c) 2026
	Neura Tech AI & Lumina AI.

	Neura Tech AI
	Official AI Research & Development Organization

	Lumina AI
	Collaborating AI Research Organization

	Project:
	Nexa AI

	Model:
	Nexa-AI-4x4B-Instruct

	This NOTICE file is provided for attribution purposes only and does not
	modify, replace, or supersede the terms of the Apache License, Version 2.0.