ModelsOfficial

Liquid AI 发布 74M 参数流式音频编解码器 LiquidCodec

Liquid AI Releases LiquidCodec, a 74M-Parameter Streaming Audio Codec

Why we picked it

Liquid AI 把 LFM3 语音模型底下的音频编解码器单独开源了出来。74M 参数、单 CPU 线程跑 361 倍实时速度,做端侧语音助手或实时通话应用的人可以省掉 GPU 编码开销,值得跟 EnCodec、Mimi 这类编解码器对比一下再选型。

Liquid AI 发布流式音频编解码器 LiquidCodec,参数量约 74M,是其 LFM3-Audio-2B 和 LFM3-Omni-2B 两个语音与多模态模型的底层音频组件。该编解码器以 12.5 Hz 的帧率运行,使用 12 个 codebook 进行离散编码,适合与自回归语言模型对接。

Image · marktechpost
Body

性能方面,LiquidCodec 在单 CPU 线程上的编码速度达到实时约 361 倍,且以流式方式工作,可以边接收音频边输出 token,不需要等待完整音频输入。低帧率加低计算开销的组合,使音频 token 序列比传统高帧率编解码器更短,能降低下游语音模型在长对话、实时通话场景下的推理成本。

LiquidCodec 定位为 LFM3 系列语音能力的基础设施,与 LFM2 系列此前强调的端侧部署方向一致。对在边缘设备或 CPU 环境下构建实时语音交互的开发者来说,单线程 361 倍实时的编码速度意味着音频前端不再是瓶颈。

Original · marktechpost

Liquid AI Releases LiquidCodec, a 74M-Parameter Streaming Audio Codec

Liquid AI released LiquidCodec, a ~74M-parameter streaming audio codec that powers LFM3-Audio-2B and LFM3-Omni-2B. It runs at 12.5 Hz with 12 codebooks and encodes about 361x faster than real time on a single CPU thread. The post Liquid AI Releases LiquidCodec, a 74M-Parameter Streaming Audio Codec appeared first on MarkTechPost .

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include LiquidCodec, Liquid AI, 音频编码, LFM3.

Primary source verification: published directly by marktechpost with first-party authority.