---
format: "aidr-story-markdown/v1"
id: "980a540185d32a5c7b6dc68fc476d4acce6d898ba247e60cc102e29f6ab7aece"
canonical_url: "https://aidr.today/980a5401?lang=en"
title: "Tencent Open Sources 1.5B AuK Speech Model Trained on 1.95 Million Hours"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-10T14:45:14.000Z"
category: "Releases"
topics: ["open-source","multimodal","llm","tencent"]
source_urls: ["https://huggingnews.com/ai/tencent-open-sources-15b-auk-speech-model-trained-on-195-million-hours-15efc442"]
summary: "A hybrid rectified-flow Transformer architecture allows for zero-shot voice cloning and audio cleanup from text prompts. Tencent"
---

# Tencent Open Sources 1\.5B AuK Speech Model Trained on 1\.95 Million Hours

> [Open the canonical story](<https://aidr.today/980a5401?lang=en>)

**Published:** 2026-09-10T14:45:14.000Z
**Category:** Releases
**Topics:** open\-source, multimodal, llm, tencent

## Summary

A hybrid rectified\-flow Transformer architecture allows for zero\-shot voice cloning and audio cleanup from text prompts\. Tencent

## Sources

- [Story source](<https://huggingnews.com/ai/tencent-open-sources-15b-auk-speech-model-trained-on-195-million-hours-15efc442>)