NVIDIA Nemotron-3 Ultra is a massive 550 billion parameter (with 55B active) mixture-of-experts model released in BF16 format. It is designed for high-quality text generation and complex reasoning. The model is hosted on Hugging Face and targets developers and enterprises needing state-of-the-art open language model capabilities.
NVIDIA Nemotron 3 Ultra 550B A55B sits in PulseGate's Foundation models & chat category. It focuses on running high-performance, large-scale language model inference for complex reasoning and generation tasks. It is built as an open-source project for developers. NVIDIA Nemotron 3 Ultra 550B A55B is open source under the Apache-2.0 license. NVIDIA Nemotron 3 Ultra 550B A55B is available on the command line and API.
Behind NVIDIA Nemotron 3 Ultra 550B A55B is NVIDIA, based in the United States, and the product first shipped in 2026. Development happens publicly on GitHub with 314 stars and 72 commits in the last 90 days. Key capabilities include mixture-of-experts, large-scale inference, and instruction following.
Latest indexed changes and source events
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 verified by the PulseGate indexer
Other apps tracked under the same category.
