
Powers faster, efficient reasoning for long-running agents
A 550B MoE frontier-intelligence open model built for long-running agents. It delivers 5x faster inference and lowers the cost of complex agentic tasks by up to 30% versus other open frontier models.Ultra excels at complex tasks like coding and deep research. Long-running agents spend their time planning, using tools, recovering from failures, and deciding what to do next.
Nemotron 3 Ultra by NVIDIA is a 550 billion parameter mixture of experts (MoE) model designed for long-running agents, offering 5x faster inference and reducing costs for complex tasks by up to 30%. It is optimized for applications in coding and deep research, enhancing efficiency in planning and decision-making processes.