NVIDIA RTX Spark Laptops Gain Nemotron 3.5 Transcribe, LFM3 Models and Agent Resource Controls
Announced the update adds an open-weight speech model for transcription, speaker labels and translation, Liquid AI's LFM3 family, and GPU resource controls for background agents.. NVIDIA says its RTX Spark Windows laptops from ASUS, Dell, HP, Lenovo, Microsoft and MSI now run Nemotron 3.5 Transcribe, an open-weight speech model that handles transcription, speaker attribution…

NVIDIA announced a wave of updates for its RTX Spark Windows laptops, pairing new open-weight models with controls that let users decide how GPU capacity is split between everyday apps and background AI agents. The systems, available from ASUS, Dell, HP, Lenovo, Microsoft and MSI, combine an NVIDIA Blackwell RTX GPU and an NVIDIA Grace CPU in one superchip with up to 128GB of unified memory and up to 1 petaflop of FP4 AI performance, which NVIDIA says puts CUDA, RTX graphics and personal AI supercomputer capabilities into thin-and-light Windows laptops.
The headline model is NVIDIA Nemotron 3.5 Transcribe, an open-weight speech model that the company says turns recorded conversations into useful information. Instead of chaining separate models for transcription, speaker labels and translation, developers can use a single promptable model that handles all three and can also answer questions about the audio. It transcribes 40 languages, including seven Indic languages, and prompts can carry product terms and formatting conventions for specialized vocabulary in healthcare, financial services and customer service. NVIDIA says the model ranks No. 1 on the Hugging Face Open ASR leaderboard and No. 2 among open-weight models on Artificial Analysis, with roughly 4.3 billion active parameters, about 80% fewer than the top open-weight model. It is available on Hugging Face now, and NVIDIA says quantized versions are on the way so developers can run Transcribe locally on RTX Spark and DGX Spark systems, Jetson AGX Thor devices, or scale up on NVIDIA Blackwell data center GPUs.
NVIDIA also announced day-zero support for Liquid AI's LFM3 family of open models, described as a new generation of Liquid Foundation Models spanning text, vision, audio, omni and decision models built on one shared backbone, with base and post-trained weights available. Liquid AI says its new 1.7-billion-parameter base model outperforms its previous 2.6-billion-parameter base model. NVIDIA lists local workflows the models support, including agents running in Hermes Agent, Pi and OpenClaw that triage inboxes, manage calendars, read documents and app screens, transcribe and translate speech, answer spoken questions about handwritten math, and route support tickets or voice requests. Developers can run LFM3 on RTX Spark through llama.cpp and vLLM and fine-tune the models locally, according to NVIDIA.
On the system side, NVIDIA App gains resource management controls that let users choose how GPU resources are shared between a foreground app and background-running agents, prioritizing the app for responsiveness or giving the agent more capacity to finish work faster. NVIDIA says the setting lives under System > Advanced in the NVIDIA App and requires hardware-accelerated GPU scheduling to be enabled in Windows Settings.
The update also covers gaming on the same hardware. NVIDIA says S-GAME's action role-playing title Phantom Blade Zero launched with ray tracing and DLSS 4.5, including DLSS Dynamic Multi Frame Generation for up to 6x frame generation or an adjustable multiplier to hit a target frame rate. NVIDIA RTX Remix 1.6, released the week before, brought the modding platform's path-traced remastering and DLSS support to Minecraft Java Edition 1.21.1.
Source: NVIDIA.
InteliToys is editorial. We do not sell NVIDIA RTX Spark. Watch the source tape in the branded lightbox below.


