{"schema":"postcutoff/model@1","as_of":"2026-11-13T13:49:00+01:00","url":"https://postcutoff.com/m/kyutai-moshi/","md":"https://postcutoff.com/m/kyutai-moshi/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"kyutai-moshi","name":"Kyutai Moshi / Hibiki-Zero (full-duplex speech models)","org":"Kyutai","family":"Moshi","released":"2024-09-17","status":"current","type":"audio/speech","modality_in":["audio"],"modality_out":["audio","text"],"open_weights":true,"model_license":"cc-by-4.0","context_window":null,"max_output":null,"knowledge_cutoff":null,"pricing":null,"price_line":"Price not published","access":[{"provider":"Hugging Face","model_id":"kyutai/moshiko-pytorch-bf16","url":"https://github.com/kyutai-labs/moshi"},{"provider":"Hugging Face","model_id":"kyutai/hibiki-zero-3b-pytorch-bf16","url":"https://huggingface.co/kyutai/hibiki-zero-3b-pytorch-bf16"},{"provider":"Web demo","url":"https://moshi.chat"}],"capabilities":[{"name":"Open full-duplex spoken dialogue","detail":"7B temporal transformer modelling user and Moshi audio streams simultaneously with an 'inner monologue' text stream; 160 ms theoretical / ~200 ms practical latency on an L4; Mimi codec (24 kHz, 12.5 Hz, 1.1 kbps).","first":true,"discovered":"launch","source":"https://github.com/kyutai-labs/moshi"},{"name":"Hibiki-Zero simultaneous speech translation","detail":"3B model (2026-02-12) translating French, Spanish, Portuguese and German speech to English in real time with voice transfer, trained without aligned data.","first":false,"discovered":"later","source":"https://kyutai.org/blog/"},{"name":"MoshiRAG","detail":"Asynchronous knowledge retrieval via a text LLM for full-duplex speech models (2026-04-30); RL post-training for interactivity (2026-06-10).","first":false,"discovered":"later","source":"https://kyutai.org/blog/"}],"entry":null,"notes":"Moshi (announced July 2024, weights + paper Sept 2024) is widely cited as the first real-time full-duplex open spoken dialogue model; NVIDIA PersonaPlex-7B (Jan 2026) is fine-tuned from Moshiko weights. Variants: moshiko (male)/moshika (female) in PyTorch bf16/int8, MLX int4/int8/bf16, Rust/Candle. Code MIT/Apache, weights CC-BY-4.0.","verified":"2026-09-29","body_md":"Sources: https://github.com/kyutai-labs/moshi , https://kyutai.org/blog/ , https://huggingface.co/kyutai/hibiki-zero-3b-pytorch-bf16","page_url":"https://postcutoff.com/m/kyutai-moshi/","standing":{"kind":"speech and audio","newer_count":106,"newer":[{"id":"nemotron-3-5-transcribe","name":"NVIDIA Nemotron 3.5 Transcribe (30B-A4B)","org":"NVIDIA","released":"2026-10-28","page_url":"https://postcutoff.com/m/nemotron-3-5-transcribe/"},{"id":"hy-simulmt-30b-a3b","name":"Hy-SimulMT-30B-A3B","org":"Tencent","released":"2026-10-22","page_url":"https://postcutoff.com/m/hy-simulmt-30b-a3b/"},{"id":"eleven-realtime-v1-max","name":"Eleven Realtime v1 Max / Mini","org":"ElevenLabs","released":"2026-10-20","page_url":"https://postcutoff.com/m/eleven-realtime-v1-max/"},{"id":"voxtral-mini-transcribe-3","name":"Voxtral Mini Transcribe 3 (+ Mini Transcribe Realtime 3)","org":"Mistral AI","released":"2026-10-20","page_url":"https://postcutoff.com/m/voxtral-mini-transcribe-3/"},{"id":"voxtral-small-transcribe-3","name":"Voxtral Small Transcribe 3","org":"Mistral AI","released":"2026-10-20","page_url":"https://postcutoff.com/m/voxtral-small-transcribe-3/"}],"before":null,"after":{"id":"kyutai-tts-stt","name":"Kyutai TTS 1.6B / Kyutai STT + Unmute","org":"Kyutai","released":"2025-07-03","page_url":"https://postcutoff.com/m/kyutai-tts-stt/"},"weights":"public","licence":"cc-by-4.0","licence_permissive":false,"summary":"Weights are public. Licence: CC BY 4.0. See the licence for commercial terms."}}