Qwen/Qwen3-TTS
Qwen3-TTS is an advanced text-to-speech model by Alibaba's Qwen team, delivering stable, expressive, and low-latency speech generation across 10 languages. Key capabilities: - 9 preset voices — Vivian, Serena, Uncle_Fu, Dylan, Eric, Ryan, Aiden, Ono_Anna, Sohee — covering diverse genders, ages, and accents - Voice cloning — clone any voice from a short (~3s) audio sample via the voice_id parameter - Instruction control — adjust tone, emotion, and speaking style with natural language (e.g. "speak slowly and calmly", "excited tone") - 10 languages — English, Chinese, Japanese, Korean, German, French, Russian, Spanish, Italian, Portuguese - Streaming support — real-time PCM streaming with ~97ms first-byte latency - Multiple output formats — WAV, MP3, FLAC, PCM Built on a 1.7B parameter architecture using discrete multi-codebook language modeling for end-to-end speech synthesis without cascading errors. Uses a custom 12Hz acoustic tokenizer that preserves paralinguistic information and environmental audio details.
Try the model with a sample request using the same API shape shown in the reference page.
Text to convert to speech
Natural language instruction for tone/emotion control (e.g. "angry tone", "whispering", "speak slowly"). (Default: empty)
Select the desired language for the speech output. Use Auto for auto-detection.
Select the desired format for the speech output. Supported formats include mp3, opus, flac, wav, and pcm.
Preset voice name (Vivian, Serena, etc.) or a voice_id from /v1/voices/add for voice cloning.
The service tier used for processing the request. When set to 'priority', the request will be processed with higher priority (only applies to models that support it).
Tweak the overall style and tone of the conversation by giving some 'master' instructions. (Default: Be a helpful assistant)
Sample from the set of tokens with highest probability such that sum of probabilies is higher than p. Lower values focus on the most probable tokens.Higher values sample more low-probability tokens (Default: 0.9, 0 < top_p ≤ 1)
repetition penalty. Value of 1 means no penalty, values greater than 1 discourage repetition, smaller than 1 encourage repetition. (Default: 1, 0.01 ≤ repetition_penalty ≤ 5)
Sample from the best k (number of) tokens. 0 means off (Default: 0, 0 ≤ top_k < 1000)
maximum length of the newly generated generated text.If explicitly set to None it will be the model's max context length minus input length or 65536, whichever is smaller (Default: empty, 1 ≤ max_new_tokens ≤ 1000000)
A unique identifier representing your end-user, which can help monitor and detect abuse. Avoid sending us any identifying information. We recommend hashing user identifiers.. (Default: empty)
Float that represents the minimum probability for a token to be considered, relative to the probability of the most likely token. Must be in [0, 1]. Set to 0 to disable this. (Default: 0, 0 ≤ min_p ≤ 1)
Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. (Default: 0, -2 ≤ presence_penalty ≤ 2)
Positive values penalize new tokens based on how many times they appear in the text so far, increasing the model's likelihood to talk about new topics. (Default: 0, -2 ≤ frequency_penalty ≤ 2)
How to format the response
temperature to use for sampling. 0 means the output is deterministic. Values greater than 1 encourage more diversity (Default: 0.7, 0 ≤ temperature ≤ 1)
You can add more items with the button
Seed for random number generator. If not provided, a random seed is used. Determinism is not guaranteed. (Default: empty, -9223372036854776000 ≤ seed < 18446744073709552000)
Adjust settings and run the demo to see the model output generated here in real-time.
