{"id":"amazon-nova-2-sonic","nom":"Amazon Nova 2 Sonic","editeur":"Amazon","pays":"États-Unis","poids_ouverts":false,"poids_selon":"Artificial Analysis","licence":null,"licence_selon":null,"liens":{"page":"https://docs.aws.amazon.com/nova/latest/nova2-userguide/using-conversational-speech.html"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/amazon-nova-2-sonic/","releve":"2026-11-09","en_bref":"19e sur 38 en raisonnement à l'écoute (Big Bench Audio) ; 19e sur 34 au délai avant le premier son.","presentation":"Modèle parole à parole d'Amazon, servi par Amazon Bedrock, pour les assistants vocaux et l'automatisation du service client, avec appel d'outils asynchrone et gestion des interruptions. Sa mise à jour de mars 2026 a réduit la latence et rendu la voix plus fidèle sur les codes, montants et adresses.","sortie":"2025-12-02","langues":{"annoncees":{"annoncees":"7 langues pour les voix expressives, dont le français","nombre":7,"francais":true,"liste":["en","fr","it","de","es","pt","hi"]},"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["presentation"],"lu":"https://docs.aws.amazon.com/nova/latest/nova2-userguide/using-conversational-speech.html (« Amazon Nova 2 Sonic enables real-time conversational AI with speech input and output. »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://docs.aws.amazon.com/nova/latest/nova2-userguide/using-conversational-speech.html (« Asynchronous tool handling that executes tool calls while maintaining conversation flow, allowing the assistant to continue speaking while tools process in the background. »), 23/10/2026"},{"appuie":["langues"],"lu":"https://docs.aws.amazon.com/nova/latest/nova2-userguide/using-conversational-speech.html (« Expressive voices are offered, including both masculine-sounding and feminine-sounding voices, in the following languages: English (US, UK, India, Australia) French Italian German Spanish Portuguese Hindi »), 23/10/2026"},{"appuie":["sortie","presentation"],"lu":"https://docs.aws.amazon.com/nova/latest/nova2-userguide/release-notes.html (« The Amazon Nova 2 Sonic refresh (March 2026 release) improves on the initial general availability launch from December 2, 2025, with the following enhancements: Better speech generation quality with higher verbatim fidelity (fewer speech generation hallucinations) for customer service call automation use cases involving alphanumeric codes, currency values, addresses, email addresses, and phone numbers. »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://docs.aws.amazon.com/nova/latest/nova2-userguide/release-notes.html (« Latency optimizations that reduce user-perceived latency at the 50th percentile (p50) by 150 ms. »), 23/10/2026"}],"classements":[{"id":"voix-a-voix/raisonnement","nom":"Raisonnement à l'écoute, Big Bench Audio","rang":19,"sur":38,"valeur":88.1,"unite":"%","chiffres":{"cout_par_heure_audio_entree":4.894854137468235,"hebergeur":"Amazon Bedrock"},"voisins":["openai-gpt-live-1","deepslate-opal"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-speech","releve":"2026-11-09"}},{"id":"voix-a-voix/delai","nom":"Délai avant le premier son","rang":19,"sur":34,"valeur":1.14,"unite":"s","chiffres":{"cout_par_heure_audio_entree":4.894854137468235,"hebergeur":"Amazon Bedrock"},"voisins":["openai-gpt-realtime-2","google-gemini-3-8-live"],"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":"Nova 2.0 Sonic (Mar 2026)"}}