nvidia

llama-3.1-nemotron-ultra-253b-v1

PREVIEW

Superior inference efficiency with highest accuracy for scientific and complex math reasoning, coding, tool calling, and instruction following.