Double click image to zoom in
Click image to open expanded view
M5stack LLM (Large Language Model) Module (AX630C) Edge AI Development Board Suitable for Offline Model Operation
No reviews
- Sale price
- $99.65
- List price
- $142.35
- You save
- $42.70 (30%)
Prices include VAT and any applicable duties.
Delivery estimate
Estimated arrival is calculated for your delivery market.
Ships in 1-3 business days. Remote areas or special items may take longer.
M5Stack Module LLM AX630C Edge AI Inference Module
M5Stack Module LLM is a compact offline AI inference module based on the AX630C SoC. It is intended for embedded voice interaction, local language-model experiments and edge AI systems that require on-device processing rather than a continuous cloud connection.

Compute Platform
- AX630C SoC with dual Cortex-A53 processors at 1.2 GHz.
- NPU performance: up to 3.2 TOPS at INT8 and 12.8 TOPS at INT4.
- 4 GB LPDDR4: 1 GB system memory plus 3 GB dedicated to hardware acceleration.
- 32 GB eMMC 5.1 on-board storage.
- Listed typical power use: 5 V / 0.5 W idle and 5 V / 1.5 W at full load.

Integrated Functions
- StackFlow framework support for edge AI development.
- Built-in KWS wake-word, ASR speech-recognition, LLM and TTS functions.
- MSM421A microphone, AW8737 audio driver and 8 ohm 1 W speaker.
- Three RGB status LEDs driven by LP5562.
- Serial communication defaults to 115200 8N1 and is adjustable.

Development and Updates
- Compatible with Arduino, UiFlow and documented StackFlow workflows.
- Firmware upgrade is available through an SD card or the Type-C port.
- The module includes a button for entering firmware-download mode.
- Listed dimensions: 54 x 54 x 13 mm; operating temperature: 0 to 40 C.

Model Compatibility
- This module uses AXERA-specific model formats; standard public model files cannot necessarily be loaded directly.
- Select model packages made for the AX630C platform and follow the official software-update workflow.
- Confirm memory, model, peripheral and host-device compatibility before planning a deployment.

