A compact text model from Tencent's HunyuanMT line, running in FP8 precision to keep memory footprint small. At 1.8B parameters it trades raw capability for efficiency, making it practical to deploy on modest hardware. Its 262K token context window is notably generous for a model this size.