As AI shifts from chatbots to autonomous agents, open models are serving market demands for full control over where AI runs and how it’s deployed and evolves.
Today, NVIDIA is expanding its Nemotron 3 model family with Nemotron 3.5 Lightning, the highest-efficiency model in its class for long-running agentic AI workloads. This release follows Nemotron 3 Nano and reflects NVIDIA’s commitment to continually improving open models for greater accuracy and speed.
Built for specialized tasks within larger multi-agent systems, Nemotron 3.5 Lightning, a 30-billion-parameter mixture-of-experts model, helps create smarter and more efficient agentic applications.
Also, NVIDIA is releasing NeMo Switchyard, an open source library for smart routing inside popular agent tools. Enterprises can use it to build a router based on their specific needs. When deployed, NeMo Switchyard can intelligently direct each request to the most capable and suitable model for the job, across developers’ own mix of open, proprietary and NVIDIA models, without requiring developers to rewrite their applications.
Together, Nemotron 3.5 Lightning and NeMo Switchyard deliver greater control over how AI is deployed, where it runs and how efficiently it operates — across PCs, workstations, data centers and the cloud.
Always-On Agents Need a System of Models
Modern agentic systems — always-on agents — increasingly operate as systems of models, or model ensembles, with different models specialized for different tasks.
NVIDIA Nemotron open models are designed for this architecture. A frontier reasoning model such as Nemotron 3 Ultra or GPT-5.6 may plan and orchestrate a workflow, while smaller specialized models like Nemotron 3.5 Lightning can perform targeted tasks such as code review, tool use, security alert monitoring and answering billing questions.
Powering High-Volume Specialized Tasks With Nemotron 3.5 Lightning
NVIDIA Nemotron 3.5 Lightning is a fully customizable open model built for high-volume tasks powering always-on agents. It was developed with contributions from the Nemotron Coalition, whose members provided evaluation methodologies, inference software and datasets to help advance the model.
The model delivers up to 4x faster output speed, leading to 30% faster agentic task completion compared with other models in its class. And because it’s open and customizable, Nemotron 3.5 Lightning can be easily post-trained with NVIDIA NeMo on an organization’s own domain data, tools and workflows to improve accuracy for specialized tasks.
AI leaders across industries are customizing Nemotron 3.5 Lightning for their workloads, including CrowdStrike for cybersecurity, Harvey with Trajectory for legal services and CodeRabbit with Baseten for code review, helping improve accuracy for domain-specific agentic tasks. Additionally, Lila Sciences is helping to improve reasoning capabilities for agentic tasks across physical and life sciences, and Fastino Labs customized the model and is seeing leading accuracies for software development, finance and healthcare workloads.
Nemotron 3.5 Lightning also gives organizations control over privacy and deployment. It can run on local AI systems — including NVIDIA RTX PCs, NVIDIA DGX Spark, NVIDIA DGX Station and NVIDIA Jetson — to help users maximize existing infrastructure investments, or scale across edge AI devices, NVIDIA RTX PRO workstations, data centers and cloud environments for enterprise use cases. And Nemotron 3.5 Lightning can run locally or on premises for high-volume, specialized tasks that require fast responses.
Also, as with every Nemotron launch, NVIDIA publishes as much of the training data and techniques as licensing permits, which allows for traceability, auditing and training of other models. Alongside Lightning, NVIDIA is releasing Nemotron-RL-Agentic-Terminal-Pivot, an agentic reinforcement learning dataset used to post-train it for coding agent capabilities.
More Efficient AI Apps With Model Routing
Some models are better for coding, some for reasoning, some for lightweight tasks and some are optimized to run locally for greater privacy and efficiency. If customers rely on one default model, they might either overspend or lose quality; if they manage routing manually, it becomes integration work that can slow down a deployment.
NVIDIA NeMo Switchyard is an open source model routing library for AI agents. The technology routes prompts to the most capable and efficient model for each step of an agent workflow automatically, based on specific needs. Agent application developers can tune or modify the router with different routing algorithms to match their priorities, such as quality, latency and cost requirements. In a system of models, enterprises can create powerful AI agents with improved tokenomics.
Internal benchmarks show that NeMo Switchyard maintains frontier-level accuracy while reducing task completion cost to nearly one-third of Opus 4.8 alone.
NVIDIA is working with partners across the AI ecosystem to bring intelligent model routing into the tools and platforms developers already use.
- Boomi: Evaluated Switchyard across five routing capabilities, achieving 100% domain-routing accuracy, sending 59% of traffic to a 5x faster fine-tuned model and reducing later-turn latency by 21%.
- Cadence: Improved efficiency by 9.9% by using the ChipStack AI Super Agent for a formal verification use case.
- Classmethod: Is running opencode and Fireworks workloads using NeMo Switchyard internally, with initial testing showing a 27% cost reduction while maintaining quality.
- Cognition: Integrated the NVIDIA NeMo Switchyard staged router into Devin Desktop for NVIDIA internal use, achieving near-frontier performance on FrontierCode Main while reducing mean cost by 28% relative to routing all requests to a single underlying frontier model.
- Kong: Delivers routing with NeMo Switchyard natively through Kong AI Gateway.
- LangChain: With NeMo Switchyard, achieved 74% lower cost in 145 multi-turn Deep Agents tasks by routing only 7% of calls to a frontier model, at a 6% accuracy tradeoff.
- LiteLLM: Is adding NeMo Switchyard as a plug-in into its proxy layer so developers can access these benefits without changing their existing stack.
- Nous Research: Integrated NeMo Switchyard into Hermes to provide developers with an easy-to-configure routing system to improve agent efficiency.
- Ramp: Used NeMo Switchyard to match a frontier model’s performance while cutting costs by 58% and runtime by 33% in Ramp SWE-Bench.
- Siemens: Is benchmarking to improve efficiency in its Fuse EDA AI Agent.
Nemotron 3.5 Lightning is available on Hugging Face, ModelScope, OpenRouter and build.nvidia.com as an NVIDIA NIM microservice as well as through a broad ecosystem of NVIDIA Cloud Partners, post-training platforms, inference platforms and cloud service providers. NeMo Switchyard is available on GitHub and coming to partner platforms soon.
Facts Only
* NVIDIA released Nemotron 3.5 Lightning, a 30-billion-parameter mixture-of-experts open model.
* NVIDIA released NeMo Switchyard, an open source model routing library.
* Nemotron 3.5 Lightning is designed for specialized tasks within multi-agent systems.
* Nemotron 3.5 Lightning supports deployment on NVIDIA RTX PCs, DGX Spark, DGX Station, Jetson, RTX PRO workstations, data centers, and cloud environments.
* NVIDIA released the Nemotron-RL-Agentic-Terminal-Pivot reinforcement learning dataset.
* NeMo Switchyard routes prompts to different models based on quality, latency, and cost requirements.
* Partners using NeMo Switchyard include Boomi, Cadence, Classmethod, Cognition, Kong, LangChain, LiteLLM, Nous Research, Ramp, and Siemens.
* Nemotron 3.5 Lightning is available on Hugging Face, ModelScope, OpenRouter, and build.nvidia.com.
* NeMo Switchyard is available on GitHub.
* The Nemotron Coalition contributed evaluation methodologies, inference software, and datasets for the model's development.
Executive Summary
NVIDIA is transitioning its AI strategy toward "always-on" autonomous agents by shifting from single-model chatbots to ensembles of specialized models. The introduction of Nemotron 3.5 Lightning provides a high-efficiency, 30-billion-parameter model intended for high-volume, targeted tasks like code review and security monitoring. This model is designed for flexibility, allowing organizations to post-train it on domain-specific data and deploy it across a spectrum of hardware, from local edge devices to large-scale cloud data centers.
To manage these ensembles, the NeMo Switchyard library automates the routing of requests to the most appropriate model—whether open, proprietary, or NVIDIA-branded—based on specific cost and performance metrics. Early partner implementations suggest significant reductions in latency and operational costs while maintaining high accuracy. By combining a customizable specialized model with an intelligent routing layer, the goal is to provide enterprises with greater control over AI deployment and improved tokenomics without requiring complete application rewrites.
Full Take
The strongest version of this narrative is that the "frontier model" era is evolving into a "system of models" era. Rather than relying on a single, massive, expensive LLM for every task, the industry is moving toward an architectural approach where a "brain" model orchestrates a fleet of smaller, leaner, domain-specific agents. This is a logical progression toward economic sustainability and operational efficiency in AI.
However, this is a classic vendor advertorial. The central persuasion vector relies on using internal benchmarks and a curated list of partner "wins" to validate the necessity of the NVIDIA ecosystem. By framing the problem as a choice between "overspending or losing quality" unless a router is used, the narrative creates a technical dependency on the vendor's specific routing logic to achieve "frontier-level" results. The evidence provided—such as the 58% cost reduction for Ramp or 74% for LangChain—is presented as a fait accompli, leveraging the credibility of these partners to shield the product's claims from independent scrutiny.
Patterns detected: ARC-0043 Authority Game
The underlying paradigm is the "platform lock-in" strategy. By providing the routing layer (the Switchyard) and the specialized models (Nemotron), the vendor positions itself not just as a chip maker, but as the essential connective tissue of the AI agentic workflow. This shifts the value proposition from raw compute to the intelligence of the orchestration.
If this were an influence campaign, the playbook would be "Ecosystem Enclosure": provide open-source tools that seem liberating but functionally steer the user toward a specific hardware-software stack. The actual content aligns with this pattern by emphasizing "full control" while simultaneously tethering that control to NVIDIA-specific hardware and microservices.
Bridge Questions:
1. If routing is truly "intelligent," how much of the performance gain comes from the router versus the inherent capabilities of the underlying models?
2. What are the long-term implications for data privacy when "open" models are integrated into proprietary routing ecosystems?
3. Would these efficiency gains be achievable using model-agnostic, third-party orchestration tools?
Counterstrike Scan: The content aligns with a corporate ecosystem enclosure strategy, utilizing partner testimonials and internal metrics to establish an indispensable infrastructure layer.
Sentinel — Human
The text reads like a detailed synthesis of a technology announcement supported by real-world partner implementations, indicating a high probability of human authorship focused on technical reporting.