Thin, Light and Built for Everything: NVIDIA RTX Spark Windows Laptops Gain New Models, Agent Tools and Games
NVIDIA RTX Spark laptops now ship with a Blackwell RTX GPU and Grace CPU unified in a single super‑chip offering up to 128 GB of memory and 1 PFLOP of FP4 AI performance, bringing personal‑AI supercomputer capabilities to thin‑and‑light Windows devices. The launch adds the open‑weight Nemotron 3.5 Transcribe speech model (~4.3 B parameters, 80 % fewer than the leading open model) which tops Hugging Face Open ASR and supports 40 languages plus speaker attribution and translation. New resource‑management controls in the NVIDIA App let users allocate GPU compute between foreground apps and background agents, and the platform debuts with games like Phantom Blade Zero (ray tracing + DLSS 4.5) and Minecraft RTX Remix 1.6. Engineers can now run high‑quality transcription locally on RTX Spark laptops or scale to DGX Spark and Blackwell data‑center GPUs.
⚡ Key Takeaways
- Nemotron 3.5 Transcribe has ~4.3 B active parameters and ranks No. 1 on Hugging Face Open ASR.
- RTX Spark laptops combine a Blackwell RTX GPU and Grace CPU in a single superchip with up to 128 GB unified memory.
- Resource‑management feature in NVIDIA App lets users prioritize foreground apps or background agents for GPU share.
- Model can be deployed via NeMo‑Speech.cpp or vLLM (NVFP4 version) and quantized builds are announced.
- The feature requires hardware‑accelerated GPU scheduling to be enabled in Windows Settings.
Engineers can embed a state‑of‑the‑art, open‑weight transcription model directly on a portable laptop without cloud latency, while the new GPU‑sharing controls ensure AI agents don’t starve interactive applications of performance.
✅ Practical Steps
- Install the Nemotron 3.5 Transcribe model from Hugging Face and run it on an RTX Spark laptop using NeMo‑Speech.cpp or vLLM.
- Enable hardware‑accelerated GPU scheduling in Windows Settings, then open NVIDIA App → System > Advanced to set GPU resource priorities between your foreground app and any background agents.
Want the full story? Read the original article.
Read on NVIDIA Blog ↗