NVIDIA announced the release of Nemotron 3.5 Super VL, a 120-billion-parameter open model, alongside two new edge models on October 15, 2026. The new additions are designed to cover the full spectrum of agent deployments from enterprise servers down to edge hardware.
New Models and Availability
NVIDIA released Nemotron 3.5 Super VL, an open 120-billion-parameter model built in collaboration with the Nemotron Coalition. According to NVIDIA, it delivers fast task completion and high accuracy for specialized always-on agents.
Joining Super VL are two new edge models: Nemotron 3.5 Bolt 74B-A9B 24GB and Nemotron 3.5 Bolt E4B 4GB. Super VL and Bolt 74B-A9B are available now, while Bolt E4B is scheduled for release on October 23.
Customization and Platform Integrations
Organizations across various industries are post-training Super VL on their own data and workflows. Companies such as Scale AI, AfterQuery, AI21, Alice, and Mercor are customizing the model for applications ranging from financial agents to coding assistants.
Software platform companies integrating or interoperating with Nemotron 3.5 Super VL include Atlassian, CrowdStrike, Dassault Systemes, Perplexity, Salesforce, and Synopsys. The model is available on Hugging Face, Open Router, ModelScope, and as an NVIDIA NIM microservice.
Key facts and where they come from
- NVIDIA released Nemotron 3.5 Super VL, an open 120-billion-parameter model.
NVIDIA today released Nemotron 3.5 Super VL, an open 120-billion-parameter model that delivers the fastest task completion and highest accuracy for specialized always-on agents.
- Super VL and Bolt 74B-A9B are available now, with Bolt E4B coming October 23.
Super VL and Bolt 74B-A9B are available now, with Bolt E4B coming Oct. 23.
- Nemotron 3.5 Super VL was trained on about 26 trillion tokens.
Nemotron 3.5 Super VL was trained on about 26 trillion tokens across pretraining and continued pretraining
