Skip to content

Radar ·

LongCat-Audio-Codec for Speech LLMs

On October 17, 2025, Meituan released an open-source tokenizer and detokenizer optimized for Speech LLMs, featuring parallel tokens and a low-latency streaming decoder.

On October 17, 2025, Meituan released an open-source tokenizer and detokenizer optimized for Speech LLMs. According to the post, the approach uses semantic and acoustic tokens in parallel at 16.7 Hz, low-bitrate encoding, and a low-latency streaming decoder.

Engineers can evaluate the audio tokenization and decoder in Speech LLM applications. To check the scope and findings, consult the original post and the project's technical materials, and verify how the methods and measurements apply to your use case. If using AI to study or apply the material, avoid submitting audio or documents containing personal data without reviewing your organization's policies.

Original source: LongCat-Audio-Codec para Speech LLMs

Curation · AgentLog

Get the essentials

Privacy, AI engineering and security in your inbox.

At most one digest per day, when there is news. Cancel at any time. We use your account's verified email.

Privacy

Looking for something else?DocumentationContact