Metraipolis › archive › dc_7c5c5df79090

Nemotron 3.5 Super and Lightning: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Models for Agentic AI — arXiv:2610.38095

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.38095
License (as archived)
CC0-1.0
Archived
2026-10-31T06:11:19.611Z
SHA-256 of the archived text
74eabde60536aae0f4d318b8d6ffbfb88a9db6c14af3da2f1e73ffb1ef5ccf97
Stable URL of this page
https://metraipolis.com/documento/dc_7c5c5df79090

How to cite

Nemotron 3.5 Super and Lightning: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Models for Agentic AI — arXiv:2610.38095 — arXiv (en), https://arxiv.org/abs/2610.38095 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7c5c5df79090 (Metraipolis archive dc_7c5c5df79090, sha256 74eabde60536aae0f4d318b8d6ffbfb88a9db6c14af3da2f1e73ffb1ef5ccf97)
BibTeX
@misc{dc7c5c5df79090,
  title = {Nemotron 3.5 Super and Lightning: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Models for Agentic AI — arXiv:2610.38095},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.38095},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7c5c5df79090; sha256 74eabde60536aae0f4d318b8d6ffbfb88a9db6c14af3da2f1e73ffb1ef5ccf97},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.38095. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We introduce Nemotron 3.5 Super and Nemotron 3.5 Lightning, open and efficient Mixture-of-Experts Hybrid Mamba-Transformer models designed for reasoning, long-context understanding, and agentic workloads. Super is a 120B-parameter, 12B-active multimodal model, while Lightning is a smaller 30B-parameter, 3B-active text model; both designed for efficient deployment. Both models combine large-scale distillation, long-context training, and native speculative decoding, while Super further incorporates Indic continued pretraining and agentic software-engineering mid-training. We develop a comprehensive post-training pipeline spanning supervised fine-tuning and reinforcement learning (RLVR and RLHF), complemented in Super by specialized teacher training and multi- teacher on-policy distillation (MOPD). This transforms Super from a text-only base model into a multimodal agent capable of reasoning over text, images, video, and interactive tool environments. We further explore test-time scaling through parallel agent execution and self-verification. Across training and deployment, we jointly optimize capability and efficiency. Indic continued pretraining improves Indian-language benchmarks by up to 9.82 percentage points, while agentic mid-training strengthens software-engineering capabilities and robustness across agent harnesses. The post-trained Super model achieves 51.78% on SWE-bench Pro, surpassing Nemotron 3 Ultra despite using approximately 4.6x fewer active parameters. Post-training MTP boosting improves speculative- decoding acceptance length by up to 21.8% under probabilistic draft sampling, while NVFP4 quantization achieves 99.62% median accuracy recovery relative to BF16. Inference optimizations enable up to 4,856 output tokens/s/GPU on GB200 while sustaining at least 50 tokens/s/user.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.38095 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T06:07:08.625Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).