NVIDIA Nemotron 3 Super Release Delivers Higher Throughput for Agentic AI

NVIDIA is launching Nemotron 3 Super—a 120-billion-parameter open model with 12 billion active parameters designed to run complex agentic AI systems at scale. 

Available now, the model combines advanced reasoning capabilities to efficiently complete tasks with high accuracy for autonomous agents.

Perplexity offers its users access to Nemotron 3 Super for search and as one of 20 orchestrated models in Computer. Companies offering software development agents such as CodeRabbit, Factory, and Greptile are integrating the model into their AI agents along with proprietary models to achieve higher accuracy at lower cost.

And life sciences and frontier AI organizations including Edison Scientific and Lila Sciences will power their agents for deep literature search, data science, and molecular understanding.

Industry leaders such as Amdocs, Palantir, Cadence, Dassault Systèmes, and Siemens are deploying and customizing the model to automate workflows in telecom, cybersecurity, semiconductor design, and manufacturing. 

Nemotron 3 Super has a 1-million-token context window, allowing agents to retain full workflow state in memory and prevent goal drift.

According to NVIDIA, Nemotron 3 Super has set new standards, claiming the top spot on Artificial Analysis for efficiency and openness with leading accuracy among models of the same size. 

The model also powers the NVIDIA AI-Q research agent to the No. 1 position on DeepResearch Bench and DeepResearch Bench II leaderboards, benchmarks that measure an AI system’s ability to conduct thorough, multistep research across large document sets while maintaining reasoning coherence. 

Nemotron 3 Super uses a hybrid mixture-of-experts (MoE) architecture that combines three major innovations to deliver higher throughput and higher accuracy than the previous Nemotron Super model. 

On the NVIDIA Blackwell platform, the model runs in NVFP4 precision. That cuts memory requirements and pushes inference up to 4x faster than FP8 on NVIDIA Hopper, with no loss in accuracy. 

NVIDIA is releasing Nemotron 3 Super with open weights under a permissive license. Developers can deploy and customize it on workstations, in data centers or in the cloud.

The model was trained on synthetic data generated using frontier reasoning models. NVIDIA is publishing the complete methodology, including over 10 trillion tokens of pre- and post-training datasets, 15 training environments for reinforcement learning and evaluation recipes. Researchers can further use the NVIDIA NeMo platform to fine-tune the model or build their own. 

Nemotron 3 Super is designed to handle complex subtasks inside a multi-agent system and has a high-accuracy tool calling that ensures autonomous agents reliably navigate massive function libraries to prevent execution errors in high-stakes environments, such as autonomous security orchestration in cybersecurity, NVIDIA said.

NVIDIA Nemotron 3 Super, part of the Nemotron 3 family, can be accessed at build.nvidia.com, Perplexity, OpenRouter and Hugging Face.

Dell Technologies is bringing the model to the Dell Enterprise Hub on Hugging Face, optimized for on-premise deployment on the Dell AI Factory, advancing multi-agent AI workflows. HPE is also bringing NVIDIA Nemotron to its agents hub to help ensure scalable enterprise adoption of agentic AI. 

For more information about this news, visit www.nvidia.com.