---
title: "Speech-Omni-Lite"
created: 2026-04-15
updated: 2026-04-15
type: entity
tags: [speech-model, multimodal, model]
sources:
  - raw/papers/2026/03/2603.09627.md
---

# Speech-Omni-Lite

Speech-Omni-Lite is a portable speech interface designed for vision-language models (VLMs), enabling lightweight speech interaction capabilities. It is referenced in the context of [[omni-modal-llm]] as part of the broader ecosystem of multimodal speech-language models. The approach focuses on efficiency and portability, making speech capabilities accessible without requiring full-scale speech LLM architectures.

## Key Facts

| Attribute | Value |
|-----------|-------|
| Type | Portable Speech Interface for VLMs |
| Focus | Lightweight, efficient speech interaction |
| Related Paper | arXiv:2603.09627 |

## Related

- [[omni-modal-llm]] — Omni-Modal LLM paradigm that Speech-Omni-Lite contributes to
- [[speech-llm]] — Broader speech LLM category

## References

- Speech-Omni-Lite: Portable Speech Interfaces for Vision-Language Models. arXiv:2603.09627.