# NVIDIA Says Nemotron 3.5 Transcribe Will Soon Run Locally on RTX Spark

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/43028
Published: 2026-10-28T22:46:05.000Z
Updated: 2026-10-28T22:46:05.000Z
Section: Technology

> The update adds Liquid AI’s LFM3 models and controls for sharing GPU capacity between foreground apps and background AI agents.

NVIDIA’s RTX Spark update adds Liquid AI models and GPU-sharing controls, while local support for its 40-language Nemotron 3.5 Transcribe speech model is still coming.

NVIDIA has made the open-weight model available, but the quantized versions needed to run it locally on RTX Spark are coming soon. It supports speech transcription, translation and speaker labeling across 40 languages. The speaker labels mark which speaker produced each portion of a transcription.

The update also adds Liquid AI’s LFM3 series to RTX Spark. The models can run locally and be fine-tuned. New GPU resource controls let users allocate computing capacity between foreground applications and background AI agents.

RTX Spark laptops from ASUS, Dell, HP, Lenovo, Microsoft and MSI start at $2,599. They offer up to 128GB of unified memory and up to 1 petaflop of FP4 AI computing performance.