Nvidia launches open models for workplace agents and local PCs
By Automate Basics, written with AI from TechCrunch's original
Nvidia has introduced an enterprise agent model and smaller models for AI PCs and workstations. The open-weight models give organizations options for running agents on their own infrastructure or locally, according to TechCrunch.
What changed
Nvidia has launched Nemotron 3.5 Super VL for enterprise agents and two smaller Nemotron Bolt models for local use, TechCrunch reports. Nvidia says Super VL is suited to tasks that involve following instructions and using tools. Organizations can adapt the open-weight model with their own data and run it on their own infrastructure.
The Bolt models serve a different setting. Nvidia says Bolt 74B-A9B 24GB is designed for local tasks involving several steps, tool use and web browsing. The smaller Bolt E4B is aimed at conversational agents. Running a model locally could matter to a team that wants an agent to work with sensitive material without sending that work to the cloud.
Who can use Nvidia's new models?
Nvidia's Super VL is intended for organizations running agents on Nvidia infrastructure, according to TechCrunch. The Bolt models are intended for compatible AI PCs and workstations, including machines using Nvidia RTX Spark and DGX Spark. Nvidia identifies enterprise knowledge workers among the audiences for those PCs, alongside consumers and programmers.
TechCrunch reports that Super VL and Bolt 74B-A9B are available now. Nvidia says Bolt E4B is due later in October. The report does not specify a subscription plan, model price or hardware cost, so teams should not assume that access to the model includes a suitable machine. Availability of a model also does not mean an organization has already approved it for workplace data.
How do I try these models at work?
A workplace team with the appropriate hardware could start with a narrow document task rather than a fully autonomous agent. For example, an accounting team could ask a model to identify fields in documents, then compare each result with the original before using it in a workflow. TechCrunch lists document work and financial accounting among the tasks Nvidia discusses for Super VL, but does not provide a setup procedure.
For a local Bolt model, a small conversational or multistep task would be a more relevant first check. Confirm whether the model runs on the intended machine, whether any connected tools behave as expected, and whether the output stays within the organization's data-handling rules. Nvidia's performance claims are not a substitute for checking results on your own work.
Frequently asked questions
What is the difference between Super VL and the Bolt models?
Nvidia positions Nemotron 3.5 Super VL for enterprise agents running on an organization's infrastructure. The Bolt models target local use on compatible AI PCs and workstations. Bolt 74B-A9B is meant for multistep tasks and tool use, while Bolt E4B is aimed at conversation, according to TechCrunch.
Can these models keep workplace tasks off the cloud?
Nvidia says organizations can run Super VL on their own infrastructure and that the Bolt models are designed for local machines. Whether a particular workplace task stays off the cloud depends on how the organization sets up the model and any tools it connects to.
Are all of Nvidia's new models available now?
TechCrunch reports that Nemotron 3.5 Super VL and Bolt 74B-A9B are available now. Nvidia says the smaller Bolt E4B is coming later in October. The report does not give prices or say which workplace plans, if any, include access.
For you
Learn to do this with AI, free
Short courses for people who aren't engineers: ChatGPT, Claude, prompting and automation. Every lesson is free, with no card.
Start a free courseFor companies
Make an AI or automation tool?
List it in the directory for $9.99, refunded in full if it isn't approved, or claim its page if it is already here.
Related reading
Laplace uses AI agents to stress-test robots and self-driving carsLaplace Intelligence uses AI agents to create difficult situations for robots and self-driving vehicles inside interactive simulations. The product is aimed at teams that need to find failures before deploying physical AI.
Google brings Gemini data analysis to Sheets on iOSGoogle says eligible users can now use Gemini in Google Sheets on iOS to ask questions about spreadsheet data, get analytical insights and create charts. The feature is rolling out gradually.
Google lets Gemini users export interactive apps to Google SitesGoogle now lets people export interactive apps from the Gemini app into private Google Sites drafts. Eligible users can edit the site, add content and publish it without copying code or rebuilding the app.
Written with AI from TechCrunch's original and published after automatic checks: every figure here appears in the original, and no sentence is copied from it. The picture is AI-generated. The original is the authority.
Get the next one by email
One email, most weeks: what changed in AI for people who use it at work, and what to do about it.
Source: Nvidia launches new open models to run agents in the enterprise and on PCs, TechCrunch, 19 Oct 2026.