Accès ouvert
2026
preprint
OpenAlex
NVIDIA, :, Aaron Blakeman, Austin Thomas et autres
We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine …
Accès ouvert
2026
preprint
OpenAlex
NVIDIA, :, Aaron Blakeman, Austin Thomas et autres
We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine …
Accès ouvert
2026
preprint
OpenAlex
Aryan Sood, Shantanu Acharya, Gaurav Kumar Nayak
Inference with large language models (LLMs) on long sequences is computationally expensive due to the quadratic complexity of self-attention. Distributed blockwise methods such as Star Attention reduce this cost by sharding context across hosts, but rely on prepending a static, content-blind copy …
Accès ouvert
2026
preprint
OpenAlex
Aryan Sood, Shantanu Acharya, Gaurav Kumar Nayak
Inference with large language models (LLMs) on long sequences is computationally expensive due to the quadratic complexity of self-attention. Distributed blockwise methods such as Star Attention reduce this cost by sharding context across hosts, but rely on prepending a static, content-blind copy …
Accès ouvert
2026
preprint
OpenAlex
NVIDIA, :, Aakshita Chandiramani, Aaron Blakeman et autres
We describe the pre-training, post-training, and quantization of Nemotron 3 Super, a 120 billion (active 12 billion) parameter hybrid Mamba-Attention Mixture-of-Experts model. Nemotron 3 Super is the first model in the Nemotron 3 family to 1) be pre-trained in NVFP4, 2) leverage …
Accès ouvert
2026
preprint
OpenAlex
NVIDIA, :, Aakshita Chandiramani, Aaron Blakeman et autres
We describe the pre-training, post-training, and quantization of Nemotron 3 Super, a 120 billion (active 12 billion) parameter hybrid Mamba-Attention Mixture-of-Experts model. Nemotron 3 Super is the first model in the Nemotron 3 family to 1) be pre-trained in NVFP4, 2) leverage …
Accès ouvert
2025
preprint
OpenAlex
NVIDIA, :, Aaron Blakeman, Aaron Grattafiori et autres
We introduce the Nemotron 3 family of models - Nano, Super, and Ultra. These models deliver strong agentic, reasoning, and conversational capabilities. The Nemotron 3 family uses a Mixture-of-Experts hybrid Mamba-Transformer architecture to provide best-in-class throughput and context lengths of up to …
Accès ouvert
2025
preprint
OpenAlex
NVIDIA, :, Aaron Blakeman, Aaron Grattafiori et autres
We introduce the Nemotron 3 family of models - Nano, Super, and Ultra. These models deliver strong agentic, reasoning, and conversational capabilities. The Nemotron 3 family uses a Mixture-of-Experts hybrid Mamba-Transformer architecture to provide best-in-class throughput and context lengths of up to …
Accès ouvert
2025
preprint
OpenAlex
NVIDIA, :, Aaron Blakeman, Aaron Grattafiori et autres
We present Nemotron 3 Nano 30B-A3B, a Mixture-of-Experts hybrid Mamba-Transformer language model. Nemotron 3 Nano was pretrained on 25 trillion text tokens, including more than 3 trillion new unique tokens over Nemotron 2, followed by supervised fine tuning and large-scale RL on …
Accès ouvert
2025
preprint
OpenAlex
NVIDIA, :, Aaron Blakeman, Aaron Grattafiori et autres
We present Nemotron 3 Nano 30B-A3B, a Mixture-of-Experts hybrid Mamba-Transformer language model. Nemotron 3 Nano was pretrained on 25 trillion text tokens, including more than 3 trillion new unique tokens over Nemotron 2, followed by supervised fine tuning and large-scale RL on …
Accès ouvert
2025
preprint
OpenAlex
Krishna C. Puvvada, Faisal Ladhak, Santiago Akle Serrano, Cheng-Ping Hsieh et autres
We present a decoder-only Transformer architecture that robustly generalizes to sequence lengths substantially longer than those seen during training. Our model, SWAN-GPT, interleaves layers without positional encodings (NoPE) and sliding-window attention layers equipped with rotary positional encodings (SWA-RoPE). Experiments demonstrate strong performance …
Accès ouvert
2025
preprint
OpenAlex
Nvidia Nvidia, Aarti Basant, Abhinav Khattar, Adithya Renduchintala et autres
As inference-time scaling becomes critical for enhanced reasoning capabilities, it is increasingly becoming important to build models that are efficient to infer. We introduce Nemotron-H, a family of 8B and 56B/47B hybrid Mamba-Transformer models designed to reduce inference cost for a given …