LaunchAI ModelsDevelopers
Jun 4, 1:02 PM
NVIDIA launches Nemotron 3 Ultra for agent reasoning
Nemotron 3 Ultra is a 550B-parameter MoE model (55B active) that achieves 5x higher throughput. In SWE-bench and Terminalbench 2.0, it reduces token use by up to 30% for agentic tasks, cutting costs.
·
Jun 4, 1:02 PM