NVIDIA Releases Nemotron 3.5 Bolt E4B 4GB: An Open Small Language Model for Local Voice Agents, Game Characters and Robots

NVIDIA has released Nemotron 3.5 Bolt E4B 4GB, an open small language model built on Gemma 4 E4B and designed to operate locally on hardware like RTX GPUs and Jetson modules. By requiring only 4GB of VRAM and offering a 128K-token context window, the model provides developers with a resource-efficient option for integrating real-time voice interaction, image analysis, and tool usage into robotics and gaming applications.