Skip to main content
Qwen Cloud home page
Docs
CosyVoice

Voice list

System voice catalog

These tables list all system voices for CosyVoice. Requirements:
  • Each model supports a specific set of voices -- you cannot mix them across models.
  • The text value must be in a language the voice supports. Unsupported languages cause pronunciation errors or unnatural output.
  • For SSML-enabled voices, pass SSML content in the text parameter. See SSML guide.
  • For Instruct-enabled voices, pass Instruct-formatted text in the instruction parameter.
  • For timestamp-enabled voices, set word_timestamp_enabled to true (enableWordTimestamp in Java SDK). Timestamp data is delivered as word-level timing events through WebSocket callbacks, not embedded in the audio data.
    • Python SDK: pass word_timestamp_enabled through additional_params: SpeechSynthesizer(model=model, voice=voice, callback=callback, additional_params={"word_timestamp_enabled": True})

cosyvoice-v3-flash

ScenarioVoice nameVoice parameterAttributeAgeLanguageSSMLInstructTimestampSample
Social companionship (Benchmark voice)Long AnyanglonganyangSunny young man20-30Chinese (Mandarin), EnglishYesYesYesSample
Social companionship (Benchmark voice)Long AnhuanlonganhuanEnergetic and cheerful female20-30Chinese (Mandarin), EnglishYesYesYesSample
Child's voice (Benchmark voice)Long Huhulonghuhu_v3Innocent and lively girl6-10Chinese (Mandarin), EnglishYesYesYesSample
DialectsLong Anyuelonganyue_v3Energetic Cantonese male25-35Chinese (Cantonese), EnglishYesNoYesSample
DialectsLong Shangelongshange_v3Authentic Northern Shaanxi male25-35Chinese (Shaanxi dialect), EnglishYesNoYesSample
DialectsLong Anminlonganmin_v3Innocent young girl18-25Chinese (Minnan dialect), EnglishYesNoYesSample
TelesalesLong Yingxiaolongyingxiao_v3Sweet-voiced saleswoman20-25Chinese (Mandarin), EnglishYesNoYesSample
Customer serviceLong Yingxunlongyingxun_v3Young and inexperienced male20-25Chinese (Mandarin), EnglishYesNoYesSample
Customer serviceLong Yingtaolongyingtao_v3Gentle and composed female25-30Chinese (Mandarin), EnglishYesNoYesSample
Voice assistantLong Anyunlonganyun_v3Homey and warm-hearted male30-35Chinese (Mandarin), EnglishYesNoYesSample
Voice assistantLong Anwenlonganwen_v3Elegant and intellectual female25-35Chinese (Mandarin), EnglishYesNoYesSample
Voice assistantLong Anlilonganli_v3Crisp and composed female25-35Chinese (Mandarin), EnglishYesNoYesSample
Voice assistantLong Anlanglonganlang_v3Fresh and crisp male20-25Chinese (Mandarin), EnglishYesNoYesSample
Voice assistantLong Yingmulongyingmu_v3Elegant and intellectual female25-30Chinese (Mandarin), EnglishYesNoYesSample
Social companionshipLong Anzhilonganzhi_v3Wise and mature young male25-35Chinese (Mandarin), EnglishYesNoYesSample
Social companionshipLong Anyalonganya_v3Elegant and classy female25-35Chinese (Mandarin), EnglishYesNoYesSample
Social companionshipLong Anqinlonganqin_v3Approachable and lively female20-25Chinese (Mandarin), EnglishYesNoYesSample
AudiobooksLong Wanjunlongwanjun_v3Delicate and soft-spoken female20-30Chinese (Mandarin), EnglishYesNoYesSample
AudiobooksLong Yichenlongyichen_v3Free-spirited and energetic male20-30Chinese (Mandarin), EnglishYesNoYesSample
AudiobooksLong Laobolonglaobo_v3World-weary old man60+Chinese (Mandarin), EnglishYesNoYesSample
AudiobooksLong Laoyilonglaoyi_v3Worldly and calm auntie60+Chinese (Mandarin), EnglishYesNoYesSample
Short video voice-overLong Jiqilongjiqi_v3Dorky and cute robot20-30Chinese (Mandarin), EnglishYesNoYesSample
Short video voice-overLong Hougelonghouge_v3Classic Monkey King20-25Chinese (Mandarin), EnglishYesNoYesSample
Short video voice-overLong Daiyulongdaiyu_v3Delicate and talented female voice15-25Chinese (Mandarin), EnglishYesNoYesSample
Livestreaming e-commerceLong Anxuanlonganxuan_v3Classic female livestreamer30-40Chinese (Mandarin), EnglishYesNoYesSample
Instruct settings The following voices support Instruct. Pass these formats in the instruction parameter:
  1. Set emotion
    • Format: "你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <情感值> with an emotion value.)
    • Example: "你说话的情感是neutral."
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  2. Set scenario and emotion
    • Format: "你正在进行<场景>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <场景> with a scenario and <情感值> with an emotion value.)
    • Example: "你正在进行闲聊互动，你说话的情感是neutral."
    • Scenarios: 闲聊互动 (casual conversation), 新闻播报 (news broadcast), 广告促销 (ad promotion), 比赛解说 (sports commentary), 一些儿童内容解说 (commentary for children's content), 语音导航 (voice navigation), 脱口秀表演 (stand-up comedy).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  3. Set role and emotion
    • Format: "你现在说话的角色是<角色>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <角色> with a role and <情感值> with an emotion value.)
    • Example: "你现在说话的角色是一个旁白，你说话的情感是neutral."
    • Roles: 一个旁白 (narrator).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  4. Set identity and emotion
    • Format: "你正在以一个<身份>的身份说话，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <身份> with an identity and <情感值> with an emotion value.)
    • Example: "你正在以一个故事机的身份说话，你说话的情感是neutral."
    • Identities: 故事机 (storytelling machine).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  1. Set emotion
    • Format: "你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <情感值> with an emotion value.)
    • Example: "你说话的情感是neutral."
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  2. Set scenario and emotion
    • Format: "你正在进行<场景>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <场景> with a scenario and <情感值> with an emotion value.)
    • Example: "你正在进行闲聊对话，你说话的情感是neutral."
    • Scenarios: 闲聊对话 (casual conversation), 比赛解说 (sports commentary), 深夜电台广播 (late-night radio broadcast), 剧情解说 (plot summary), 诗歌朗诵 (poetry reading), 科普知识推广 (science popularization), 产品推广 (product promotion), 脱口秀表演 (stand-up comedy).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  3. Set role and emotion
    • Format: "你现在说话的角色是<角色>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <角色> with a role and <情感值> with an emotion value.)
    • Example: "你说话的角色是温和客服，你说话的情感是neutral."
    • Roles: 温和客服 (gentle customer service).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  1. Set emotion
    • Format: "你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <情感值> with an emotion value.)
    • Example: "你说话的情感是neutral."
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  2. Set scenario and emotion
    • Format: "你正在进行<场景>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <场景> with a scenario and <情感值> with an emotion value.)
    • Example: "你正在进行自由对话，你说话的情感是neutral."
    • Scenarios: 自由对话 (free conversation), 广告促销 (ad promotion).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  3. Set role and emotion
    • Format: "你现在说话的角色是<角色>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <角色> with a role and <情感值> with an emotion value.)
    • Example: "你说话的角色是傲娇公主，你说话的情感是neutral."
    • Roles: 傲娇公主 (tsundere princess), 元气少女 (energetic girl), 可爱孩童 (cute child), 机器人 (robot), 小猪佩奇 (Peppa Pig).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  4. Set identity and emotion
    • Format: "你正在以一个<身份>的身份说话，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <身份> with an identity and <情感值> with an emotion value.)
    • Example: "你正在以一个故事机的身份说话，你说话的情感是neutral."
    • Identities: 故事机 (storytelling machine), 儿童玩具 (children's toy).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.

cosyvoice-v3-plus

ScenarioVoice nameVoice parameterAttributeAgeLanguageSSMLInstructTimestampSample
Social companionship (Benchmark voice)Long AnyanglonganyangSunny young man20-30Chinese (Mandarin), EnglishYesYesYesSample
Social companionship (Benchmark voice)Long AnhuanlonganhuanEnergetic and cheerful female20-30Chinese (Mandarin), EnglishYesYesYesSample
Instruct settings The following voices support Instruct. Pass these formats in the instruction parameter:
  1. Set emotion
    • Format: "你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <情感值> with an emotion value.)
    • Example: "你说话的情感是neutral."
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  2. Set scenario and emotion
    • Format: "你正在进行<场景>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <场景> with a scenario and <情感值> with an emotion value.)
    • Example: "你正在进行闲聊互动，你说话的情感是neutral."
    • Scenarios: 闲聊互动 (casual conversation), 新闻播报 (news broadcast), 广告促销 (ad promotion), 比赛解说 (sports commentary), 一些儿童内容解说 (commentary for children's content), 语音导航 (voice navigation), 脱口秀表演 (stand-up comedy).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  3. Set role and emotion
    • Format: "你现在说话的角色是<角色>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <角色> with a role and <情感值> with an emotion value.)
    • Example: "你现在说话的角色是一个旁白，你说话的情感是neutral."
    • Roles: 一个旁白 (narrator).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  4. Set identity and emotion
    • Format: "你正在以一个<身份>的身份说话，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <身份> with an identity and <情感值> with an emotion value.)
    • Example: "你正在以一个故事机的身份说话，你说话的情感是neutral."
    • Identities: 故事机 (storytelling machine).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  1. Set emotion
    • Format: "你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <情感值> with an emotion value.)
    • Example: "你说话的情感是neutral."
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  2. Set scenario and emotion
    • Format: "你正在进行<场景>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <场景> with a scenario and <情感值> with an emotion value.)
    • Example: "你正在进行闲聊互动，你说话的情感是neutral."
    • Scenarios: 闲聊对话 (casual conversation), 比赛解说 (sports commentary), 深夜电台广播 (late-night radio broadcast), 诗歌朗诵 (poetry reading), 科普知识推广 (science popularization), 产品推广 (product promotion), 脱口秀表演 (stand-up comedy).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.
  3. Set role and emotion
    • Format: "你现在说话的角色是<角色>，你说话的情感是<情感值>。" (Use Chinese. Include "。". Replace <角色> with a role and <情感值> with an emotion value.)
    • Example: "你说话的角色是温和客服，你说话的情感是neutral."
    • Roles: 温和客服 (gentle customer service).
    • Emotions: neutral, fearful, angry, sad, surprised, happy, disgusted.