LFM2.5-Audio-1.5B
Liquid AI · 2026-01-05 · 1.6B parameters
LFM2.5-Audio-1.5B is the audio model of Liquid AI's January 5, 2026 LFM2.5 launch, an end-to-end speech-conversation model with an LFM2.5 backbone, a 115M FastConformer audio encoder and an RQ-transformer generating Mimi audio tokens, plus a new LFM-based audio detokenizer. Its main weights hold 1,470,308,488 BF16 parameters; the repository also ships Kyutai's Mimi codec (96,151,393) and the detokenizer (78,532,290), for 1,644,992,171 in the released checkpoint. The card gives a 32,768-token context and English support.