Nvidia Releases New Open-Weight AI Models for Enterprise Data Centers and Local PCs
Nvidia says its 120-billion-parameter Nemotron 3.5 Super VL matches the larger Nemotron 3 Ultra on instruction-following and tool use, while completing specialized agentic tasks faster than models from Anthropic and OpenAI. Two Bolt models target local RTX Spark PCs.
Nvidia launched a new family of open-weight AI models on Thursday, October 19, 2026, targeting two distinct markets: large enterprises running agents in their own data centers, and consumers using AI locally on personal computers.
The flagship release is the Nemotron 3.5 Super VL, a 120-billion-parameter model built for enterprise agentic workloads. Nvidia says it delivers accuracy comparable to the larger Nemotron 3 Ultra on instruction-following and tool-use tasks. The company claims Super VL completes agentic tasks faster than competing models from Anthropic and OpenAI — including Opus 5, GPT 6.1 Sol, Luna, and Astra — when post-trained on specific tasks such as document intelligence, financial accounting, scientific experimentation, medical insurance claims, and security attack resilience. Nvidia lists Skydance, EY, Ramp, CrowdStrike, Perplexity, and Salesforce among companies already using the model.
Alongside Super VL, Nvidia introduced two lightweight Bolt models designed to run locally on AI PCs powered by the Nvidia RTX Spark chip and workstations running the DGX Spark. The Nemotron 3.5 Bolt 74B-A9B 24GB is a 74-billion-parameter model built to handle multistep reasoning, tool calling, and web browsing. The smaller Bolt E4B targets conversational agents. Both are aimed at consumers, programmers, and enterprise knowledge workers using hardware from partners including Microsoft and Dell.
Nvidia’s vice president of Open Models and Software, Kari Briski, said the open approach gives organizations the ability to post-train models on their own data, deploy on their own infrastructure, and retain ownership of what they build — something closed frontier models do not allow. For PC users, running models locally means delegating tasks “without sending those workloads to the cloud,” Briski wrote.
The move reflects Nvidia’s effort to serve enterprises seeking data-center independence from proprietary AI providers, while also building a market for its PC chips by positioning local AI agents as a privacy-friendly alternative to cloud-based assistants.
Super VL and Bolt 74B-A9B are available now. Bolt E4B is expected later in October 2026.