Canonical announced that NVIDIA’s new Nemotron 3.5 Lightning AI, a 30-billion-parameter hybrid Mixture-of-Experts model, is now available on Ubuntu, as users can install it easily with a ready-to-use inference Snap and just one command.
If you haven’t kept up with every new AI model, Nemotron 3.5 Lightning is NVIDIA’s compact, open AI model built for the repetitive tasks that autonomous AI agents handle.
What sets it apart, however, is that rather than using a large reasoning model for every small job, it handles tasks like tool calls, result checks, output formatting, and other frequent tasks where speed and efficiency matter more than deep reasoning.
Canonical has made it much easier to deploy the model on Ubuntu. Instead of setting up and configuring the inference software yourself, you can install Canonical’s pre-packaged inference runtime with sudo snap install nemotron-3-5-lightning. The Nemotron package is available at launch and works on compatible NVIDIA-powered workstations, servers, and edge devices.
The model supports a context window of up to one million tokens, which helps long-running AI agents keep information during complex, multi-step tasks. Canonical also points out that the Snap package offers a standardized runtime, secure execution, verified distribution, automatic updates, and less maintenance.
Nemotron 3.5 Lightning is also fully customizable. NVIDIA gives developers the model weights, training data, and recipes, so they can fine-tune it for specific tasks instead of only using it as a remote service. NVIDIA says the model can run locally, in the cloud, or in data centers.
Moreover, the model can be up to four times faster than similar models for some tasks, as it is designed for always-on agents, so simple operations do not have to go through a much larger model, which would slow things down and use more resources.
For additional details, see the official Canonical announcement.
