Models · MarkTechPost ·
NVIDIA AI Releases Nemotron 3 Ultra: An Open 550B Mixture-of-Experts Hybrid Mamba-Transformer for Long-Running Agents
NVIDIA has released Nemotron 3 Ultra, an open 550B-parameter mixture-of-experts hybrid Mamba-Transformer model for long-running agents. The model uses 55B active parameters, supports up to 1M-token context, and is reported to offer up to about 6x higher inference throughput than comparable open LLMs at similar accuracy