NVIDIA has launched Nemotron 3 Ultra, a new enterprise-focused AI model designed to improve business automation, reasoning, and large-scale AI deployments across industries. The model focuses on multi-step reasoning, planning, and self-correction. It is designed to support AI agents capable of managing sophisticated workflows with minimal human involvement.
NVIDIA CEO Jensen Huang introduced Nemotron 3 Ultra, which is a 550-billion-parameter open-weight AI model. He introduced it during his Computex 2026 keynote in Taipei. The chip-maker takes a major step deeper into enterprise AI software and autonomous agent development with this launch.
The model uses a mixture-of-experts architecture with approximately 55 billion active parameters per token and 90% sparsity. It delivers significantly higher efficiency than its total parameter count suggests.
Nemotron 3 Ultra outperformed all US-based open-weights models, including Google’s Gemma 4 31B, in its initial tests. NVIDIA also stated that the model generates more than 300 output tokens per second. In comparison, competing models such as DeepSeek and Moonshot typically delivers between 50 and 100 tokens per second. Furthermore, the company claims the model lowers costs by roughly 30% for complex agentic workloads.
Also Read: NVIDIA Unveils RTX Spark: Driving the Shift to Autonomous AI Agent PCs
Alongside Nemotron 3 Ultra, NVIDIA launched the broader Nemotron 3 family. The lineup includes the mid-tier Super model and Nano Omni, a lightweight multimodal model built for edge devices. Moreover, Nano Omni combines vision, audio, and language capabilities to power on-device AI agents.
The company also showcased the Vera CPU, a processor designed specifically for agentic AI workloads. According to NVIDIA, the new chip's efficiency is twice that of the standard x86 server chips. On the other hand, the RTX Spark combines an Arm CPU with the Blackwell GPU and supports 128GB of unified memory.
With the rising need for AI among enterprises, there have been efforts to design AI systems specific to businesses. The development of such AI systems would provide more efficient, safe, and precise systems than those currently available.