IT Home On October 17, according to news, JetBrains announced yesterday (October 16) that its local AI programming tool, Junie Local, is now fully adapted for the NVIDIA RTX Spark platform at Day 0. Additionally, its support for the Windows system has been upgraded from a trial version to a stable version.
According to a blog post cited by IT Home, users can have AI read code, modify files, and run tests on computers that meet the requirements. The official statement says the tool requires no registration or subscription, does not charge for tokens, has no usage quota, and the relevant code does not need to be uploaded to the cloud.
In terms of performance, JetBrains said that in preliminary tests on the NVIDIA RTX Spark platform, the generation speed can reach up to 55 word elements per second.
In terms of hardware requirements, JetBrains has expanded the hardware options. In addition to RTX Spark devices, computers with at least 24GB of video memory and RTX graphics cards using the Ampere or later architecture can also run, including RTX 3090, 4090, and 5090 with 32GB of video memory. The system must use Windows 11 of the corresponding architecture, install NVIDIA drivers version 581 or later, and reserve approximately 20GB of disk space.
The new installation program automatically checks the graphics card and drivers, deploys models, and adjusts the context capacity based on available GPU memory, allowing AI to retain more code and conversation records. Users can choose between Qwen3.6-27B or the fusion model provided by JetBrains.