Truffle-1 is an AI inference engine designed to run opensource models at home, on 60 Watts.

• Run Mixtral at 22+ tokens/s
• Run Mistral at 50+ tokens/s
• Run models up to 100B parameters
• Connect via WiFi, BLE or USB-C
• Nvidia Orin iGPU
• Custom carrier board
• 200 GB/s memory bandwidth