Get Zlio on your home screen
One tap, full screen, like an app.
Claims low-power on-device LLM inference, key to bringing models to robots and edge devices.
Korean fabless firm DeepX taped out a second-generation edge AI accelerator that it says delivers a 6.7x performance gain with LPDDR5x processing-in-memory, running large LLMs on device under 5 W for robotics and other edge markets.