Skip to main content

AICore AX-M1

High-Performance M.2 AI Acceleration Module

AICore AX-M1

Radxa AICore AX-M1

Radxa AICore AX-M1

High-Performance M.2 AI Acceleration Module

logo

High-Performance M.2 AI Acceleration Module

front view

The Radxa AICore AX-M1 is a high-performance M.2 computing acceleration module designed for edge AI computing and inference applications.

Powered by AXera AX8850 SoC

Integrated Octa-core Cortex A55 CPU

Powerful AI Computing Power

Built-in acceleration for Transformer-based model inference

24 TOPS

Built-in VPU:

Supports H.264/H.265 8K@30fps encoding and decoding
Supports decoding of 16 channels of 1080p@30fps
Supports real-time encoding of 7680 * 4320@30fps + 1080p@30fps

Standardized Interface

Uses standard M.2 2280 M Key form factor,compatible with mainstream industrial and embedded motherboards

Flexible Host Adaptation

Flexible Host Adaptation

Compatible Host Systems

Compatible Host Systems
open source initiative

Supported Model

Large Language Models:

  1. DeepSeek-R1-Distill
  2. Qwen2.5
  3. Qwen3
  4. MiniCPM4
  5. SmolLM3
  6. Llama3.2
  7. Gemma2
  8. Phi3

Vision Large Models:

  1. InternVL3
  2. Qwen2.5-VL
  3. SmolVLM2
  4. CLIP
  5. YOLOWorldv2

Speech Models:

  1. Whisper
  2. SenseVoice
  3. MeloTTS

Text-to-image Generation Model:

  1. Stable Diffusion v1.5

Vision Model:

  1. YOLO Family
  2. Depth-Anything-V2
  3. Real-ESRGAN
  4. MixFormerV2
  5. LivePortrait