Post-Cutoff
Search For AI agents

KyutaiMoshi familyCurrentSpeech and audioReleased 17 Sep 2024

Kyutai Moshi / Hibiki-Zero (full-duplex speech models)

Model idkyutai/moshiko-pytorch-bf16

PricePrice not published

Checked29 September 2026

Where it stands

106newer speech and audio models since its release

The 5 newest are listed. All speech and audio models

After it from Kyutai
Kyutai TTS 1.6B / Kyutai STT + Unmute3 Jul 2025
Weights
Weights are public. Licence: CC BY 4.0. See the licence for commercial terms.

How to call it

3 ways to use Kyutai Moshi / Hibiki-Zero (full-duplex speech models), with the model id each provider expects.

Providers, model ids, endpoints and docs for Kyutai Moshi / Hibiki-Zero (full-duplex speech models)
ProviderModel idEndpointDocs
Hugging Face kyutai/moshiko-pytorch-bf16 github.com
Hugging Face kyutai/hibiki-zero-3b-pytorch-bf16 huggingface.co
Web demo moshi.chat

Pricing

Price not published.

No price source linked. Checked 29 September 2026.

Specs

Input
Audio
Output
Audio and text
Context window
Not published
Max output
Not published
Open weights
Yes
Licence
CC BY 4.0
Released
17 September 2024
Status
Current
Type
Speech and audio

What stands out

What its maker marketed at launch, and what people found later, each with its source.

  • Open full-duplex spoken dialogue

    First

    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).

    Source: github.com

  • Hibiki-Zero simultaneous speech translation

    Found after launch

    3B model (2026-02-12) translating French, Spanish, Portuguese and German speech to English in real time with voice transfer, trained without aligned data.

    Source: kyutai.org

  • MoshiRAG

    Found after launch

    Asynchronous knowledge retrieval via a text LLM for full-duplex speech models (2026-04-30); RL post-training for interactivity (2026-06-10).

    Source: kyutai.org

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.

Sources: https://github.com/kyutai-labs/moshi , https://kyutai.org/blog/ , https://huggingface.co/kyutai/hibiki-zero-3b-pytorch-bf16