NVIDIA RTX Spark Mini AI Workstation Launches, Supports Running 120-Billion-Parameter Large Language Models Locally
NVIDIA announced the official launch of compact desktop computers powered by the RTX Spark platform, supporting the local operation of large language models with up to 120 billion parameters. The devices feature a 20-core Grace CPU and a Blackwell-architecture-based RTX GPU, with up to 128GB of unified memory and AI performance of up to 1 petaflop; ASUS, Dell, and MSI have already launched related products.
Why read this
AI score 66
The device specifications and local model limits are relatively detailed, providing a basis for assessing how desktop AI workstations affect reliance on cloud APIs and token costs.
Topics
Body · AI translation
Article Body EM_StockImg_Start
EM_StockImg_End
Nvidia announced that compact desktop computers powered by the RTX Spark platform are now officially available. Asus, Dell, and MSI have already launched related products, with offerings from Acer, Gigabyte, HP, and other manufacturers to follow. The devices feature a 20-core Nvidia Grace CPU, an RTX GPU based on the Blackwell architecture, up to 128GB of unified memory, and as much as 1 Petaflop of AI computing power, supporting the local operation of large language models with 120 billion parameters. Nvidia also said that running models locally can reduce reliance on cloud APIs and token costs for related tasks.
(Source: CLS) Mid-Article Embedded Advertisement
AI Trends