NVIDIA released a set of open models that developers can download and run on their own hardware. The lineup includes Nemotron 3.5 Super VL and Bolt for text‑image tasks, Kumo for structured business data, Cosmos for robotics simulation, Alpamayo 2 Edge for autonomous‑vehicle reasoning, BioNeMo for protein design and Ising for quantum computing. All models are posted on Hugging Face, GitHub and NVIDIA’s build.nvidia.com platform, with Nemotron 3.5 Bolt slated for release on Oct 23.
Why it matters
Enterprises can now fine‑tune NVIDIA’s open models on local data without sending it to the cloud, lowering cost and keeping sensitive information in‑house.
