{"rewrite":{"id":"r_cec4035abc7432366fef0d98","clusterId":"c_6235878e92afadb4e19e762d","slug":"mistral-ai-releases-voxtral-3-voice-suite-with-cloning-and-real-time-transcription","model":"deepseek-v4-1-flash","headline":"Mistral AI Releases Voxtral 3 Voice Suite With Cloning And Real-Time Transcription","summary":"Mistral AI has released Voxtral 3, a suite of four voice models: Voxtral Small Transcribe 3 and Voxtral Mini Transcribe 3 for transcription in 50+ languages, Voxtral Mini Transcribe Realtime 3 for real-time work at 240 milliseconds of latency, and Voxtral Speech 3 for synthesis in 35+ languages and 50+ dialects. Thirty minutes of reference audio can clone a specific person's voice.","whyItMatters":"A single API now covers the whole audio chain from live transcription to cloning a named speaker from half an hour of reference audio, with Mistral pricing real-time transcription at $0.008 per minute.","webCardHtml":"\u003cp\u003eThe four models sit under one name and one API. Small carries the highest transcription accuracy of the Transcribe models, a naming choice Mistral notes is confusing next to Mini, and Mistral\u0026#39;s comparison graph puts both above the prior Voxtral Mini Transcribe 2 on error rate. The real-time model transcribes more accurately than GPT Live Transcribe at 240 milliseconds of latency, with accuracy improving as latency rises.\u003c/p\u003e\u003cp\u003eSpeech 3 reads proper nouns and numbers at a cost Mistral compares to Gemini 3.8 Flash TTS. Cloning needs 30 minutes of reference audio. Pricing runs from $0.0035 to $0.008 per minute for the transcription models, and $13 per million characters for synthesis.\u003c/p\u003e\u003cul\u003e\u003cli\u003e\u003cstrong\u003eVoxtral Small Transcribe 3\u003c/strong\u003e: High-performance transcription, 50+ languages, $0.0040 per minute\u003c/li\u003e\u003cli\u003e\u003cstrong\u003eVoxtral Mini Transcribe 3\u003c/strong\u003e: Cost-efficient transcription, $0.0035 per minute\u003c/li\u003e\u003cli\u003e\u003cstrong\u003eVoxtral Mini Transcribe Realtime 3\u003c/strong\u003e: Real-time transcription at 240 ms latency, $0.008 per minute\u003c/li\u003e\u003cli\u003e\u003cstrong\u003eVoxtral Speech 3\u003c/strong\u003e: Speech synthesis, 35+ languages and 50+ dialects, $13 per million characters\u003c/li\u003e\u003c/ul\u003e","blueskyPost":"Voxtral 3 splits into four models, and only the Realtime variant carries the 240ms latency figure. Cloning needs 30 minutes of reference audio, so the suite's most sensitive feature runs on data users must supply first.","twitterPost":"Mistral AI's Voxtral 3 clones a voice from 30 minutes of reference audio. The cloning feature depends on audio the user supplies, not on the model alone.","threadsPost":"Voxtral 3 ships as four separate models, and the 240 milliseconds of latency belongs to just one of them, Voxtral Mini Transcribe Realtime 3. Real-time work and voice cloning are different products here, not one capability.","newsletterBlurb":"Mistral AI released Voxtral 3, four voice models sharing one API: two transcription models, a real-time variant, and a synthesis model that can clone a specific voice from 30 minutes of reference audio. Mistral prices the real-time model at $0.008 per minute and synthesis at $13 per million characters.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20261021-voxtral-3/\",\"title\":\"Voxtral 3: A Suite of Voice AIs Enabling Real-Time Transcription and Voice Cloning of Specific Individuals\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":4762,"outputTokens":745,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1792581732,"createdAt":"2026-10-21T11:17:17.000Z","publishedAt":"2026-10-21T11:21:16.000Z","updatedAt":"2026-10-21T11:21:16.000Z"},"cluster":{"id":"c_6235878e92afadb4e19e762d","canonicalTitle":"リアルタイム文字起こしや特定の人物の音声クローンが可能な音声AI群「Voxtral 3」","representativeArticleId":"a_5a306ed70e9f39e8bb31a319","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Voxtral 3\"],\"studios\":[\"Mistral AI\"],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-10-21T09:39:00.000Z","lastSeenAt":"2026-10-21T09:39:00.000Z","updatedAt":"2026-10-21T11:21:19.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20261021-voxtral-3/","title":"リアルタイム文字起こしや特定の人物の音声クローンが可能な音声AI群「Voxtral 3」"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Voxtral 3"],"studios":["Mistral AI"],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":null}