NVIDIA ships Nemotron 3.5 Transcribe open speech model and new RTX Spark laptops
NVIDIA Blog · rss · 2026-10-28
NVIDIA rolled out a wave of updates for its RTX Spark thin-and-light laptops (Blackwell GPU + Grace CPU, up to 128GB unified memory, 1 petaflop FP4), now sold by ASUS, Dell, HP, Lenovo, Microsoft and MSI starting at $2,599.
- Nemotron 3.5 Transcribe: an open-weight speech model handling transcription, speaker labeling, translation and audio Q&A in one promptable model, covering 40 languages including seven Indic ones. With only 4.3B active parameters (80% fewer than the leading open-weight model), it ranks No.1 on Hugging Face Open ASR. Weights available in BF16, FP8 and NVFP4, runnable on RTX Spark, DGX Spark, Jetson AGX Thor and RTX 5090.
- GPU resource management in NVIDIA App lets users split compute between foreground apps and background agents.
- Liquid AI LFM3: open-weight model family (text, vision, audio, omni, decision) with native llama.cpp/vLLM support; the new 1.7B base model beats its previous 2.6B one.
- Gaming: Phantom Blade Zero launches with ray tracing and DLSS 4.5; RTX Remix 1.6 brings path tracing to Minecraft.
More from Embodied
- Neuralink's 72-hour home sprint lets an ALS patient call, text and use Grok by thought — elonmusk · 2026-10-30
- Musk: Neuralink aims to reroute brain signals to let paralyzed people walk within years — elonmusk · 2026-10-30
- SFO Sign Reading 'No Stopping Except Authorized Autonomous Vehicles' Goes Viral — a__tomala · 2026-10-30
- Noninvasive BCI lets unresponsive ICU patients communicate via brainwaves — ZeroStateReflex · 2026-10-30
- Musk says Neuralink's Blindsight is engineering-ready, awaiting regulatory approval — elonmusk · 2026-10-30
- Cute wheeled desktop robot serves drinks at LongJourney VC demo night — aniketmaurya · 2026-10-30