Nemotron 3.5 Bolt E4B 4GB: An Open and Efficient Small Language Model for Fast Conversational Agents

We introduce Nemotron 3.5 Bolt E4B 4GB, an open and efficient small language model further post-trained from Gemma-4-E4B-it and optimized for speed and intelligence per GB of memory used for fast conversational agents. The model carries forward Gemma's strong conversational capabilities and further improves multi-turn tool use, instruction following, and input grounding for fast, local voice/device assistants, interactive personas, and embedded/Edge interfaces. The model combines natural conversation with multi-turn agentic capabilities to power voice/device assistants, interactive characters and robotic/IoT interfaces running on GeForce RTX, Jetson, and RTX/DGX Spark. Experiments on 21 benchmarks show that Nemotron 3.5 Bolt E4B 4GB achieves the highest average accuracy while generating the fewest output tokens in both reasoning and non-reasoning modes, compared with Gemma-4-E4B-it and Qwen3.5-4B. The model is released with a DFlash drafter. With DFlash, Bolt E4B achieves $2.74\times$ the per-user decode speed of inference without speculative decoding at concurrency~1, and $1.21\times$ the generation throughput of Gemma-4-E4B-it with MTP at batch size 1.

Authors

Publication Details

Journal
arXiv (Cornell University)
Published
2026-11-16
DOI
https://doi.org/10.48550/arxiv.2611.20771
Primary Topic
Topic Modeling
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Nemotron 3.5 Bolt E4B 4GB: An Open and Efficient Small Language Model for Fast Conversational Agents

Ashwath Aithal, L. Dykes, Aishwarya Chhabra, Oluwatobi Olabiyi et al.
arXiv (Cornell University)
Topic Modeling
preprint

Nemotron 3.5 Bolt E4B 4GB: An Open and Efficient Small Language Model for Fast Conversational Agents

Ashwath Aithal, L. Dykes, Aishwarya Chhabra, Oluwatobi Olabiyi, Michael Hasper, Jian Leong, Besmira Nushi, Giedrius Burachas, Leili Tavabi, Felix Friedmann, Aditya Malte, Grzegorz Chlebus, Shaona Ghosh, Bhavesh Pawar, Manoj Kilaru, Annamalai Chockalingam, Andrii Skliar, Ambrish Dantrey, Ruvaid Virk, Oleksii Kuchaiev, Prasoon Varshney, Eileen Long, Jiaqi Zeng, Alex Hodson, Krzysztof Pawelec, Ameya Sunil Mahabaleshwarkar, Saeid Motiian, Bryan Catanzaro, Haifeng Qian, Erick Galinkin, Ruby Ye, Nancy Agarwal, Hayley Ross, David Mosallanezhad, Ryan Hill, Pablo Ribalta, J. Vialard, NVIDIA, Payal Motwani, Seph Mard, Kari Briski, Dina Yared, Thinh Le, Guillermo Siman, Amy Shen, Jonathan Cohen, Gorkem Batmaz, Katherine Luna, Ran Zilberstein, Thanh Do, Terry Kong, Chris Alexiuk, Negar Habibi, Meredith Price, Pinky Xu, Chantal Hwang, Richard Mazzarese, Micah Schaffer, Michael Ross Boone, Michael Fukuyama, Sudeep Sabnis, Pritam Biswas, Rameshwar Shivbhakta, Joey Conway, Rita Fernandes Neves, Reka Kovacs, Lily Lee, Adi Margolin, Puhui Meng, Ehsan Hosseini Asl, Rochelle Woods, Shai Ben Nun, Jose Vergara, Maka Dong, Linda Liu, Chenhan Yu, Khalil Ben Khaled, Conner Currie, Kevin Ning, Sharad Yadav, Irina Filatova, Rishabh Garg, Jian Zhang, Suhas Hariharapura Sheshadri, Kedar Pandurang Potdar, Matthew Lee, Ike Nnoli, Huiying Li, Anusha Pant, Hao Guo, Iris Liu, Ashwin Kumar, Hrishith Thadicherla, Ji-Eun Lee, Bryan Yu, Cynthia Gao, Maria Bagdasarova, Liwei Ding, Hui Wang
preprint en

Abstract

We introduce Nemotron 3.5 Bolt E4B 4GB, an open and efficient small language model further post-trained from Gemma-4-E4B-it and optimized for speed and intelligence per GB of memory used for fast conversational agents. The model carries forward Gemma's strong conversational capabilities and further improves multi-turn tool use, instruction following, and input grounding for fast, local voice/device assistants, interactive personas, and embedded/Edge interfaces. The model combines natural conversation with multi-turn agentic capabilities to power voice/device assistants, interactive characters and robotic/IoT interfaces running on GeForce RTX, Jetson, and RTX/DGX Spark. Experiments on 21 benchmarks show that Nemotron 3.5 Bolt E4B 4GB achieves the highest average accuracy while generating the fewest output tokens in both reasoning and non-reasoning modes, compared with Gemma-4-E4B-it and Qwen3.5-4B. The model is released with a DFlash drafter. With DFlash, Bolt E4B achieves $2.74\times$ the per-user decode speed of inference without speculative decoding at concurrency~1, and $1.21\times$ the generation throughput of Gemma-4-E4B-it with MTP at batch size 1.

arXiv (Cornell University)
Topic Modeling
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon