Liquid AI releases LiquidCodec, an open 74M-parameter streaming speech codec

Nov 13, 2026
The codec behind Liquid's LFM3 audio models replaces Kyutai's Mimi and runs hundreds of times faster than real time on one CPU thread.

Liquid AI has released LiquidCodec, the roughly 74 million parameter streaming speech codec behind its LFM3-Audio-2B and LFM3-Omni-2B models, with weights on Hugging Face, MarkTechPost reports. The codec turns speech into tokens a language model can predict and converts them back to audio.

It splits each 80 millisecond frame of 24 kHz audio into 12 tokens that can be predicted in parallel, and replaces Kyutai's Mimi codec used in Liquid's earlier audio model. It was trained on 264,000 hours of speech in 107 languages. Liquid AI reports encoding at about 361 times real time and decoding at about 87 times on a single CPU thread of an Apple M4 Max.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Daily AI Brief.

Also, consider following us on social media:

Free newsletter

Daily AI Brief

Daily report covering major AI developments and industry news, with both top stories and complete market updates

Industry analysis

2025 Global Business Services Agenda: Gen AI Takes Center Stage

The Hackett Group

This industry analysis by The Hackett Group explores the transformative impact of generative artificial intelligence (Gen AI) on global business services (GBS) in 2025. The study highlights the shift from exploration to acceleration of Gen AI initiatives, with 89% of executives advancing these projects to improve customer satisfaction, innovate products, and reduce costs. The report also discusses the challenges and strategies for successful Gen AI adoption, emphasizing the need for a technology-enabled operating model and the importance of reskilling the workforce.

Read more
Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.