cd /news/artificial-intelligence/nvidia-debuts-three-new-open-ai-mode… · home › topics › artificial-intelligence › article
[ARTICLE · art-150708] src=ca.finance.yahoo.com ↗ pub= topic=artificial-intelligence verified=true sentiment=↑ positive

Nvidia debuts three new open AI models

Nvidia released three new open-weight AI models — Nemotron 3.5 Super VL, Nemotron 3.5 Bolt 74B-A9B 24GB, and Nemotron Bolt E4B 4GB — designed to run AI agents from data center systems down to desktop computers. Nvidia said Super VL targets agents handling tasks that can take several days or weeks and can be post-trained by enterprises on their own data, while Bolt 74B handles general agentic tasks such as navigating web browsers and using apps and Bolt E4B targets physical and conversational agents. The release extends Nvidia's open-model strategy, which CEO Jensen Huang has pushed as a way to increase competition against closed-weight systems and against China's open-model lead via Moonshot AI's Kimi K3 and Alibaba Cloud's Qwen.

by read2 min views6 publishedOct 15, 2026
Nvidia debuts three new open AI models
Image: Ca (auto-discovered)

Nvidia (NVDA) is expanding its family of open AI models with its latest Nemotron offerings.  The chip giant has made championing open-weight models, downloadable models that users can modify and inspect on their own, a key part of its growth strategy.

Open models are less expensive to use, giving more organizations the option to use AI services. And the more companies that use AI, the more chips Nvidia can sell in the long run.

The company's newest models — Nemotron 3.5 Super VL, Nemotron 3.5 Bolt 74B-A9B 24GB, and Nemotron Bolt E4B 4GB — are designed to run AI agents on everything from data center systems to desktop computers.

Super VL is meant to power agents running across tasks that can take several days or weeks, Nvidia explained in a release. The company says that enterprises can also post-train the model on their own data, better customizing it for their exact needs.

Bolt 74B is designed for general agentic tasks, such as navigating web browsers and using apps, while Bolt E4B is designed for physical and conversational agents.

While open-weight models generally trail closed models in terms of frontier capabilities, most users don't need such intense firepower for the jobs they need to accomplish. What's more, if a company can run open-weight models on its own hardware, they're more or less free to use. That also offers the dual benefit of lower costs and increased privacy, since companies running their own hardware don't have to upload their data to third-party servers.

China currently dominates the conversation on open models, with companies including Moonshot AI, Alibaba Cloud (BABA), and others offering popular model families such as Kimi K3 and Qwen.

Nvidia CEO Jensen Huang has pushed for US companies to develop and adopt more open-weight models to increase competition in the space and to avoid becoming locked into a system that relies on a small number of closed-weight models.

Email Daniel Howley at dhowley@yahoofinance.com. Follow him on X at @DanielHowley*.*

Click here for the latest technology news that will impact the stock market**.**

Read the latest financial and business news from Yahoo Finance

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @nvidia 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/nvidia-debuts-three-…] indexed:0 read:2min 2026-10-15 · —