One-line summary
On October 15, 2026, NVIDIA released Nemotron 3.5 Super VL, an open 120-billion-parameter model for specialized always-on agents that the company says delivers the fastest task completion and highest accuracy, alongside edge models Nemotron 3.5 Bolt 74B-A9B (24GB, available now) and Bolt E4B (arriving October 23); the three open models span large enterprise servers, edge devices and local hardware.
Event Overview
According to NVIDIA's official blog and newsroom on October 15, 2026, NVIDIA released Nemotron 3.5 Super VL that day: an open 120-billion-parameter model for specialized always-on agents, with the company saying it delivers the fastest task completion and highest accuracy. Two new edge models joined it: Nemotron 3.5 Bolt 74B-A9B (24GB) and Bolt E4B. Together the three open models cover the full range of agent deployments, from large enterprise servers down to edge devices and local hardware.
Key Information
| Item | Detail |
|---|---|
| Publisher / date | NVIDIA, October 15, 2026 |
| Nemotron 3.5 Super VL | Open 120B-parameter model for always-on agents; fastest task completion and highest accuracy per NVIDIA |
| Nemotron 3.5 Bolt 74B-A9B | 24GB edge model, available now |
| Nemotron 3.5 Bolt E4B | Available October 23, 2026 |
| Coverage | Large enterprise servers, edge devices, local hardware |
| Deployment | Open models; organizations can pick the tier that fits their NVIDIA infrastructure |