New NVIDIA Nemotron 3 Super Delivers 5x Higher Throughput for Agentic AI
5x throughput jump. NVIDIA's Nemotron 3 Super is built for agents—and Perplexity is already shipping it.

Why it matters
NVIDIA releases a production-ready open model optimized for agentic AI workloads, signaling a shift toward smaller, inference-efficient models that can handle real-world agent deployments at scale.
The key facts
7 to knowNemotron 3 Super: 120B parameters, 12B active parameters
5x higher throughput vs. prior generation
Designed for agentic AI systems
Open model (available now)
Perplexity integrating for user access
Advanced reasoning + task completion capabilities
Launched March 11, 2026
Go to the source
NVIDIA Blogblogs.nvidia.com
Publisher excerpt: Launched today, NVIDIA Nemotron 3 Super is 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…