{"id":"kyutai-moshi","nom":"Moshi","editeur":"Kyutai","pays":"France","poids_ouverts":true,"poids_selon":null,"licence":"CC BY 4.0","licence_selon":null,"liens":{"poids":"https://huggingface.co/kyutai/moshiko-pytorch-bf16","page":"https://github.com/kyutai-labs/moshi","essayer":"https://moshi.chat/"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/kyutai-moshi/","releve":"2026-11-09","en_bref":"38e sur 38 en raisonnement à l'écoute (Big Bench Audio) ; 27e sur 31 en tenue de la conversation (Full Duplex Bench).","presentation":"Modèle de dialogue parlé en duplex intégral du laboratoire parisien Kyutai, à poids ouverts, qui traite en parallèle sa propre parole et celle de l'utilisateur, sans tours de parole imposés. Il vise la conversation naturelle à faible latence et ne sait pas appeler d'outils.","langues":{"annoncees":{"annoncees":"anglais","nombre":1,"francais":false,"liste":["en"]},"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["presentation"],"lu":"https://huggingface.co/kyutai/moshiko-pytorch-bf16/raw/main/README.md (« Moshi is a speech-text foundation model that casts spoken dialogue as speech-to-speech generation. »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/kyutai/moshiko-pytorch-bf16/raw/main/README.md (« This allows for the removal of explicit speaker turns, and the modeling of arbitrary conversational dynamics. »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/kyutai/moshiko-pytorch-bf16/raw/main/README.md (« However, the model has limited abilities for complex tasks and cannot access tools, but rather focues on natural, low-latency interactions. »), 23/10/2026"},{"appuie":["langues"],"lu":"https://huggingface.co/kyutai/moshiko-pytorch-bf16/raw/main/README.md (« **Language(s) (NLP):** English »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://raw.githubusercontent.com/kyutai-labs/moshi/main/README.md (« All models are released under the CC-BY 4.0 license. »), 23/10/2026"}],"classements":[{"id":"voix-a-voix/raisonnement","nom":"Raisonnement à l'écoute, Big Bench Audio","rang":38,"sur":38,"valeur":4.4,"unite":"%","chiffres":{"hebergeur":"Kyutai"},"voisins":["cofe-ai-flm-audio"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-speech","releve":"2026-11-09"}},{"id":"voix-a-voix/conversation","nom":"Tenue de la conversation, Full Duplex Bench","rang":27,"sur":31,"valeur":60.97,"unite":"%","chiffres":{"hebergeur":"Kyutai"},"voisins":["cofe-ai-flm-audio","vita-freeze-omni"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-speech","releve":"2026-11-09"}}],"noms_dans_les_sources":{"aa_voix_a_voix":"Moshi"}}