NVIDIA releases open Nemotron 3.5 Bolt E4B 4GB SLM, top accuracy with fewest tokens

kalyan_kpl · x · 2026-11-18

NVIDIA open-sourced Nemotron 3.5 Bolt E4B 4GB, a small language model further post-trained from Gemma-4-E4B-it and optimized for speed and intelligence in fast conversational agents, targeting 4GB-class deployment. NVIDIA reports it 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. Paper on arXiv (2611.20771).

Original post →

More from Models

Models channel →