# Introducing NVIDIA Nemotron 3 Nano Omni: Long-Context Multimodal Intelligence for Documents, Audio and Video Agents

Canonical URL: https://www.traeai.com/articles/2b2e49e0-10fa-406f-84ea-53141faa24a5
Original source: https://huggingface.co/blog/nvidia/nemotron-3-nano-omni-multimodal-intelligence
Source name: Hugging Face Blog
Content type: article
Language: 英文
Score: 9.2
Reading time: 13 分钟
Published: 2026-04-28T15:58:57+00:00
Tags: NVIDIA, 多模态, 模型, Hugging Face

## Summary

NVIDIA 推出 Nemotron 3 Nano Omni，支持文本、图像、视频和音频的多模态理解，性能领先多个复杂任务基准。

## Key Takeaways

- Nemotron 3 Nano Omni 在文档、语音、视频等多模态任务中达到顶级精度。
- 采用 Mamba-Transformer 混合架构，支持长上下文和高吞吐量推理。
- 提供 BF16、FP8 等多种模型权重下载，适合实际工程部署。

## Citation Guidance

When citing this item, prefer the canonical traeai article URL for the AI-readable summary and include the original source URL when discussing the underlying source material.