Year: 2026
Type: preprint
Abstract: We introduce Nemotron-SEA-LION-v4.8, a family of Southeast Asian Languages In One Network (SEA-LION) models built upon NVIDIA Nemotron 3. The family includes 30B-A3B and 120B-A12B models, with both continued-pretrained base checkpoints and post-trained variants. We adapt the models using Southeast Asian, reasoning, code, and multilingual parallel data, followed by post-training with supervised fine-tuning and online on-policy distillation. On SEA-HELM, the 30B-A3B model improves the overall SEA score from 46.06 to 51.57, while the 120B-A12B model improves from 49.30 to 63.44. Across seven Southeast Asian languages, we observe broad capability gains with the 120B-A12B model showing broader and more consistent improvements across tasks.
Source: arXiv (Cornell University)
Authors: Aulia, Adila, Dabeer, Ahmed, Jeongmi, Ahn, Sajeban, Antonyrex, Adwin, Chan Hok Teng, Nicholas, Cheng Zi Yi, Esther, Choa Hsueh Mei, Jonathan, Heng, Montalan, Jann Railey Estrada, Ren, Lee Chwan, Yi, Leong Wai, Qi, Leong Wei, Rachel, Liew, Peerat, Limkonchotiwat, Mokhtar, Muhammad Ridzuan Bin, Karthik, Nagarajan, Raymond, Ng Boon Cheong, Tai, Ngee Chia, Gang, Ngui Jian, Ngan, Nguyen Thanh, David, Ong Tat-Wee, Mark, Pereira, Benjamin, Phang Shi Wei, Joseph, Poon, Hamsawardhini, Rengarajan, Yosephine, Susanto, Anocha, Sutaveephamochanon, Meng, Tan Choon, Evelyn, Tan Chor Phin, Sheryl, Tan Le Min, Jessica, Tan Siao Wei, Yixian, Tan, Panuthep, Tasawong, Yun, Tee Jun, Walter, Teng Kok Wai, Leslie, Teo Eng Sipp, William, Tjhi, Pume, Tuchinda, Donghang, Wu, Xianbin, Yong, Zhou, Zhang
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7213530745
Cites:
Cited by:
Related to: