Skip to content
TechCrunch·

🤖Nvidia Launches Nemotron 3.5 Super VL and Bolt Models for Enterprise and Local AI Tasks

Nvidia's new models promise faster, more efficient AI tasks

TL;DR

Nvidia introduces Nemotron 3.5 Super VL and Bolt models for enterprise and local AI tasks, promising faster performance and efficiency. Super VL, with 120B parameters, targets enterprise agents, while Bolt models aim for local agentic tasks on AI PCs and workstations.

Nvidia just launched Nemotron 3.5 Super VL and Bolt models, designed for enterprise and local AI tasks. The Super VL, with 120 billion parameters, promises similar accuracy to its predecessor but uses fewer resources. Bolt models, aimed at local agentic tasks, run on Nvidia RTX Spark chips and DGX Spark. If you're working on document intelligence, financial accounting, or security, these models could be a game changer. Super VL completes tasks faster than Anthropic and OpenAI models when post-trained, and it's already in use by companies like Skydance and EY. Bolt E4B, aimed at conversational agents, will be available later this month.

Nvidia Launches Nemotron 3.5 Super VL and Bolt Models for Enterprise and Local AI Tasks — TechCrunch

Key Points

1

Nemotron 3.5 Super VL, with 120 billion parameters, targets enterprise agents, available now.

2

Bolt models, designed for local agentic tasks, run on Nvidia RTX Spark chips and DGX Spark.

3

Super VL completes tasks faster than Anthropic and OpenAI models when post-trained on specific tasks.

4

Super VL is already in use by companies like Skydance, EY, Ramp, Crowdstrike, Perplexity, and Salesforce.

5

Bolt E4B, aimed at conversational agents, will be available later this month.

Why It Matters

If you're working on enterprise AI tasks or local agentic tasks on AI PCs and workstations, Nvidia's new models could significantly improve your workflow. The Super VL model, with 120 billion parameters, offers similar accuracy to its predecessor but uses fewer resources, making it more efficient. The Bolt models, designed for local agentic tasks, run on Nvidia RTX Spark chips and DGX Spark, allowing users to delegate everyday projects without sending workloads to the cloud.

nvidiaai-modelsenterprise-ailocal-agentsbolt-models

Comments

Subscribe to join the conversation...

Be the first to comment

Enjoyed this article?

Get it daily. 7am. Free. Reads in 5 minutes.

Join 3,569 builders reading daily.

Also get