Nvidia Launches Local AI Speech Model, RTX Spark Adds Support for New Models and Agent Tools

October 29: Nvidia announced that its Nemotron 3.5 Transcribe open-weight speech model now supports RTX Spark, enabling speech transcription, speaker identification and translation across 40 languages. Liquid AI's LFM3 series models are also supported, and can run locally and be fine-tuned. In addition, Nvidia has added a GPU resource management feature that allows users to allocate GPU compute between foreground applications and background AI agents. RTX Spark laptops are now available from vendors including ASUS, Dell, HP, Lenovo, Microsoft and MSI, with prices starting at $2,599, up to 128GB of unified memory, and FP4 AI compute of up to 1 petaflop.