Latest SOTA models supported on Qualcomm NPU.
AI & ML interests
On Device AI Deployment and Research
Recent Activity
View all activity
Organization Card
Welcome to Nexa AI org on HuggingFace!
Nexa AI is an on device AI deployment and research company. We craft optimized foundation models and on-device inference framework that runs any model on any device, across any backend—within minutes. Our mission is to make on device AI friction‑free and production‑ready.
On this page you’ll find
- Our own trained checkpoints
- Hand‑picked community models in GGUF or MLX formats, ready to run on nexa-sdk
Resources
- ⚙️ Download nexaSDK – get up and run models locally in minutes
- 💬 Discord Community
- 💼 Slack Community
Language Models that takes vision input and/or audio input, hand picked by Nexa Team.
-
NexaAI/gemma-3n-E4B-it-4bit-MLX
Image-Text-to-Text • Updated • 85 • 1 -
NexaAI/Qwen2.5-VL-7B-Instruct-4bit-MLX
Image-Text-to-Text • 2B • Updated • 70 -
NexaAI/SmolVLM-500M-Instruct-8bit-MLX
Image-Text-to-Text • 0.7B • Updated • 28 -
NexaAI/SmolVLM-Instruct-8bit-MLX
Image-Text-to-Text • 0.7B • Updated • 36
Latest SOTA models supported on Qualcomm NPU.
Language Models that takes vision input and/or audio input, hand picked by Nexa Team.
-
NexaAI/gemma-3n-E4B-it-4bit-MLX
Image-Text-to-Text • Updated • 85 • 1 -
NexaAI/Qwen2.5-VL-7B-Instruct-4bit-MLX
Image-Text-to-Text • 2B • Updated • 70 -
NexaAI/SmolVLM-500M-Instruct-8bit-MLX
Image-Text-to-Text • 0.7B • Updated • 28 -
NexaAI/SmolVLM-Instruct-8bit-MLX
Image-Text-to-Text • 0.7B • Updated • 36
spaces
5
Running
64
Nexa Omni Demo
🎧
Generate text from audio input
Running
79
Omnivlm Dpo Demo
👁
Ask questions about images and get detailed answers
Running
on
CPU Upgrade
29
Open LLM Leaderboard for domains
📊
Ranking for Open-sourced LLMs in different domains
Running
on
CPU Upgrade
32
Nexa AI GGUF Convertor
⚡
Submit a model for quantization and receive an email notification
models
44
NexaAI/jan-v1-4b-npu
Updated
•
3
NexaAI/Qwen3-4B-Thinking-2507-npu
Updated
•
4
NexaAI/Qwen3-4B-Instruct-2507-npu
Updated
•
3
NexaAI/OmniNeural-4B
Updated
•
109
•
101
NexaAI/qwen3-1.7B-npu
Updated
•
24
•
18
NexaAI/yolov12-npu
Updated
•
8
•
17
NexaAI/paddleocr-npu
Updated
•
9
•
15
NexaAI/qwen3-4B-npu
Updated
•
24
•
16
NexaAI/OmniVLM-968M
0.5B
•
Updated
•
1.49k
•
522
NexaAI/Qwen3-4B-GGUF
Text Generation
•
4B
•
Updated
•
420