Nvidia releases Nemotron 3.5 Super VL and Bolt open-weight models for agents
Nvidia introduced Nemotron 3.5 Super VL, a 120-billion-parameter open-weight model aimed at enterprise agent deployments, alongside two smaller Bolt models built to run locally on new RTX Spark and DGX Spark hardware. Nvidia says Super VL matches the accuracy of its larger Nemotron 3 Ultra model while completing agentic tasks faster than rival models from Anthropic and OpenAI, and lists companies including Skydance, EY, Ramp, Crowdstrike, Perplexity and Salesforce as users.
GoKawiil's interpretation of the reporting above, not reported fact.
The release expands Nvidia's open-model lineup across both large enterprise deployments and edge devices, potentially letting customers run agentic AI entirely on Nvidia hardware rather than relying on proprietary APIs. Nvidia's performance claims against Anthropic and OpenAI, companies it also invests in and supplies, highlight the increasingly tangled competitive relationships among major AI players. The named enterprise adopters suggest real-world traction, though independent verification of Nvidia's benchmark claims is not provided.
- Nemotron 3.5 Super VL is a 120B-parameter open-weight model for enterprise agents.
- Two lightweight Bolt models target local inference on new RTX Spark and DGX Spark PCs.
- Nvidia claims faster task completion than Anthropic and OpenAI models, citing adopters like Salesforce and Perplexity.
NVIDIA DGX Spark — The DGX Spark is exactly the kind of desktop AI workstation mentioned in the article, built to run Nvidia's new Bolt models locally for agent and edge computing workloads. It's a great way for developers and enthusiasts to experiment with on-device AI without relying on cloud infrastructure.
See NVIDIA DGX Spark on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: techcrunch.com — Julie Bort, 2026-10-19
Published there as: “Nvidia launches new open models to run agents in the enterprise and on PCs”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.