TLDRocket
Sign in

Nemotron 3.5 Lightning

Model ● Covered in 10 stories + Follow

Nemotron 3.5 Lightning is an open Mixture-of-Experts model from NVIDIA described in coverage as a 30B model with 3B active parameters and a 1M-token context window, positioned for “always-on” agent execution. Recent reports say it has been released alongside NeMo Switchyard, an open-source model router for routing each agent step to a suitable model, and has been made available on Amazon SageMaker JumpStart for easier deployment. Across the coverage, Nemotron 3.5 Lightning has been used for post-training and fine-tuning: CoreWeave reported faster RL-based post-training using reinforcement learning rollouts, and Nvidia/Palantir reported fine-tuning a 30B Nemotron 3.5 Lightning for supply-chain tasks. It has also been mentioned as a potential future on-device component in Perplexity AI’s Portable Computer on Nvidia-powered desktop setups.

Updated 14 October 2026

Specifications

No specifications recorded yet.

Latest developments

CoreWeave targets GPU utilization in continuous AI post-training SiliconANGLE · 1 week ago · 23
Nvidia and Palantir fine-tune a 30B Nemotron model for Nvidia’s supply chain. It beats a model 18 times its size. The New Stack · 1 month ago · 28
Perplexity AI launches Portable Computer on-device AI agent SiliconANGLE · 1 month ago · 30
The Sequence Frontier Learning - Issue 917: Understanding DeepSeek V4-Pro, GLM-5.3, NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard TheSequence · 2 months ago · 52
NVIDIA Nemotron 3.5 Lightning now available in Amazon SageMaker JumpStart Amazon Web Services · 2 months ago · 33
The Sequence Chat - Issue 912: NVIDIA’s Chris Alexiuk Talks About Nemotron, GPUs and Agentic AI Substack · 2 months ago · 6 NVIDIA AI Releases Nemotron 3.5 Lightning: A 30B Open MoE with 3B Active Parameters, and NeMo Switchyard Model Router MarkTechPost · 2 months ago · 10
NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard Deliver Faster, Smarter, More Efficient Agentic AI NVIDIA · 2 months ago · 13

Timeline

Month Quarter Year

2026

CoreWeave launched CoreWeave Forge, a full-stack platform connecting agent deployment, evaluation, and reinforcement-learning-based post-training improvements Product launch

Perplexity released Portable Computer, a locally running version of its Computer agent on supported NVIDIA hardware Deployment

DeepSeek released DeepSeek V4-Pro for general availability (GA) Model release

NVIDIA released Nemotron 3.5 Lightning, a new open mixture-of-experts agent model, along with NeMo Switchyard for model routing Model release

Nvidia released Nemotron 3.5 Lightning and NeMo Switchyard for faster agentic AI and open model routing Model release

Relationships

Products & technology

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.