{"id":"smallest-ai-lightning-v3-1","nom":"Lightning v3.1","editeur":"Smallest AI","pays":"États-Unis","poids_ouverts":false,"poids_selon":null,"licence":null,"licence_selon":null,"liens":{"page":"https://docs.smallest.ai/model-cards/text-to-speech/lightning-v-3-1","essayer":"https://docs.smallest.ai/models/getting-started"},"fiche_auto":true,"page":"https://quelleia.com/voix/modeles/smallest-ai-lightning-v3-1/","releve":"2026-10-19","en_bref":"77e sur 96 en synthèse vocale en anglais (Artificial Analysis), dans un groupe de 5 modèles trop proches pour être départagés.","presentation":"Modèle de synthèse vocale en temps réel de Smallest AI, qui produit de l'audio à 44,1 kHz avec un premier octet en environ 200 ms et propose 217 voix ainsi que le clonage instantané à partir de 5 à 15 secondes d'audio. Il change de langue en cours de phrase entre les langues dotées de voix entraînées.","langues":{"annoncees":{"annoncees":"20 codes de langue acceptés, dont 12 avec des voix entraînées ; le français passe par des voix anglaises ou hindi","nombre":20,"francais":true,"liste":["en","hi","ta","es","kn","mr","te","or","pa","ml","gu","bn","fr","de","it","nl","sv","pt","pl","ru"]},"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["presentation"],"lu":"https://docs.smallest.ai/model-cards/text-to-speech/lightning-v-3-1.md (« Real-time text to speech at 44.1 kHz: 217 voices across 12 languages, instant voice cloning, 200 ms to first byte. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://docs.smallest.ai/model-cards/text-to-speech/lightning-v-3-1.md (« Instant voice cloning with just 5-15 seconds of audio, via API or console. »), 17/10/2026"},{"appuie":["langues","presentation"],"lu":"https://docs.smallest.ai/model-cards/text-to-speech/lightning-v-3-1.md (« 20 accepted language codes plus `auto` routing; automatic identification and code-switching mid-utterance across the 12 languages with trained voices. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://docs.smallest.ai/model-cards/text-to-speech/lightning-v-3-1.md (« The model accepts these ISO 639-1 codes but has no dedicated trained voices - pronunciation and normalization are handled through the English or Hindi voice you pick. »), 17/10/2026"},{"appuie":["langues"],"lu":"https://docs.smallest.ai/model-cards/text-to-speech/lightning-v-3-1.md (« | French | `fr` | »), 17/10/2026"}],"classements":[{"id":"synthese/artificial-analysis","nom":"Synthèse vocale en anglais, Artificial Analysis","rang":77,"sur":96,"valeur":1023.86,"unite":"Elo","bas":1010.86,"haut":1036.86,"votes":2281,"chiffres":{"prix_par_million_caracteres":25,"date_sortie":"2026-03-24"},"voisins":["chatterbox","krafton-raon-speechlm"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/text-to-speech/model-families/smallestai","releve":"2026-10-19"}}],"noms_dans_les_sources":{"artificial_analysis":"Lightning v3.1"}}