Fun-ASR-Realtime released: a single model supports 30 languages and 16 dialects with leading recognition accuracy
Original title:🔥Fun-ASR-Realtime发布:语音识别再升级!
AI summary
Tongyi Lab released the Fun-ASR-Realtime real-time speech recognition model. A single model covers 30 languages and 16 dialects, with a focus on East and Southeast Asia. It achieved 87.8% semantic accuracy on the industrial dialect benchmark inhouse, far ahead of others, with local dialects approaching human level. It introduces context understanding and dynamic hotword injection to disambiguate homophones, brand names and similar terms. Streaming recognition first-character latency is kept at the hundred-millisecond level, accuracy is close to offline levels, and multilingual switching is seamless. The API is now available on the Alibaba Cloud Bailian platform.
Key Takeaways
- Tongyi Lab released the Fun-ASR-Realtime real-time speech recognition model.
- A single model covers 30 languages and 16 dialects,with a focus on East and Southeast Asia.
- It achieved 87.
- 8% semantic accuracy on the industrial dialect benchmark inhouse,far ahead of others,with local…
- It introduces context understanding and dynamic hotword injection to disambiguate homophones,brand…
Quick Facts
- Published
- Updated
- Source
- 公众号:通义实验室(千问)
- Category
- Model Releases
- AI score
- 73
Fun-ASR Realtime把多语言方言和流式一体化做得很扎实,特别是对东亚东南亚的优化,准确率提升明显,做语音应用的产品人可以直接关注。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Fun-ASR-Realtime released: a single model supports 30 languages and 16 dialects with leading recognition accuracy. July 6, 2026. https://new.nat6.net/en/news/cmr8tzbde00jjsllsayhpgjqg/