Qwen-Audio-3.0-TTS languages: 16, and Sume's language field

Qwen-Audio-3.0-TTS supports 16 languages: Arabic, Chinese, English, French, German, Indonesian, Italian, Japanese, Korean, Malay, Portuguese, Russian, Spanish, Tagalog, Thai and Vietnamese. On Sume TTS you pick the language with the language field, and the docs say to set it for every non-English transcript.

Qwen facts are from the Tongyi Lab post, read 2026-10-01. Sume facts are from the OpenAPI spec for POST /v1/tts-1.0/generate. This page does not claim that Sume covers the same list.

Which 16 languages does Qwen-Audio-3.0-TTS list?

The post lists them as written below. It credits the release date of 2026-07-20 to marktechpost.com coverage.

How do you choose a language on Sume TTS?

The language field takes a BCP-47 or ISO-639 code such as ko, ja or en, between 2 and 16 characters. If you omit it, the provider defaults to English. Sume infers ko or ja from a transcript that is only Hangul or kana, but that is a fallback, so set the field yourself.

The spec also says never to translate a non-English request into English before sending it.

What else limits a transcript?

A transcript is capped at 20000 characters, and spaces and punctuation count. For longer scripts, split the text and join the audio; combine audio files without gaps shows how.

How should you check a language before shipping?

Generate a short sample in the target language, listen to it, and confirm the voice speaks it as intended. For a worked non-English case, see Hinglish text to speech. Results come back through the job flow in Jobs and results.