Alibaba’s Tongyi Lab has launched Qwen-Audio-3.0-TTS, a hosted text-to-speech system available in Flash and Plus tiers across 16 languages.
The release is delivered exclusively through Alibaba Cloud Model Studio. Developers cannot download the model weights. The system targets two distinct workflows: real-time interaction and high-fidelity generation.
Two variants, one lineage
The Flash tier prioritises speed, achieving a first-packet latency of roughly 300 ms. The Plus tier focuses on naturalness and timbre, sacrificing speed for quality.
Both versions use the model IDs qwen-audio-3.0-tts-flash and qwen-audio-3.0-tts-plus. They communicate via a bidirectional WebSocket streaming protocol. The API outputs PCM, WAV, MP3, and Opus files at sample rates up to 48 kHz.
Features include streaming input and output, voice cloning, Voice Design, and instruction control. Alibaba provides DashScope SDKs alongside raw WebSocket examples in Python, Java, Go, C#, PHP, and Node.js. The service operates in Singapore and Beijing regions.
How the model is built
The architecture relies on two specific design choices.
First, a 12.5 Hz low-frame-rate speech tokenizer reduces autoregressive decoding costs. Fewer tokens per second of audio lowers inference latency while retaining content and speaker information.
Second, a five-stage progressive training paradigm coordinates the language model and flow-matching components. The stages include independent pretraining, joint training with data annealing, language model reinforcement learning, flow-matching robustness training, and flow-matching reinforcement learning. The team reports this pipeline improves content consistency, prosodic naturalness, voice fidelity, perceptual quality, and robustness.
The system handles one-pass long-form synthesis up to three minutes, difficult text-normalisation cases, and vocoder super-resolution for 48 kHz output.
Multilingual coverage across 16 languages
Qwen-Audio-3.0-TTS supports Arabic, Chinese, English, French, German, Indonesian, Italian, Japanese, Korean, Malay, Portuguese, Russian, Spanish, Tagalog, Thai, and Vietnamese. Seven of these languages are new additions compared to the previous line. It also covers 20 Chinese dialect regions.
On multilingual intelligibility, the model family records the best word and character error rate in 10 of the 16 languages. Flash delivers the lowest average WER/CER at 3.87. Plus is close at 3.96. Lower is better on this metric.
On speaker similarity, Plus ranks first across all 16 languages with an average of 82.75. Flash follows at 80.44. The release adds a curated preset voice library spanning the 16 supported languages, allowing teams to ship a voice without cloning one first.
Fine-grained tags for non-verbal details
For precise control, the research team embeds inline tags directly in the target text. The release adds 86 fine-grained inline tags for localized control at the phrase and word level. These cover expressive transitions and non-verbal events such as laughter, breathing, coughing, and sighing.
The Model Studio documentation splits these into two groups.
- Control tags such as
[excited],[sad],[whispers], and[asmr]set an emotion or style until the next tag. - Rich-language tags such as
[laughing],[gasp], and[clears throat]insert a single vocal effect without changing surrounding tone.
A worked example: [excited]What a beautiful day today![laughing]Let's go out and have fun together!
One limitation is worth noting. These emotion and rich-language tags are supported only in unidirectional streaming mode.
Where it stands on the leaderboard
Qwen-Audio-3.0-TTS-Plus took the top quality spot on the Artificial Analysis Speech Arena for Provider Voices. It posts an Elo near 1,236, narrowly ahead of Simba 3.2 at 1,234, and clear of Gemini 3.1 Flash TTS (1,214) and Sonic 3.5 (1,207). The lead over Simba 3.2 sits inside overlapping confidence intervals, so it is a statistical tie at the very top.
Two trade-offs are worth stating plainly. Throughput is modest: Plus generates about 16 characters per second, below Simba 3.2 (30.2), Gemini 3.1 Flash TTS (27), and Sonic 3.5 (120). Price is competitive: the listed rate is $27.59 per 1M characters, roughly a third of what ElevenLabs and MiniMax charge for the tiers it outranks. Rank and price move often, so confirm both before planning around them.
How developers and the community are reacting
The early reception is cautiously enthusiastic. The most-shared story is that a non-Western TTS topped the arena at a fraction of incumbent pricing. The most common reservations are that the model is hosted-only, that its throughput trails rivals, and that its name overlaps with the open Qwen3-TTS line.
The dashboard below aggregates that early signal across X, Reddit, and Hacker News.
What it means
Teams building voice applications now have a hosted alternative that does not require downloading weights. The Flash tier allows for immediate audio feedback in chat interfaces, while the Plus tier offers a quality benchmark that currently beats major competitors on independent leaderboards. The pricing is significantly lower than established players, though developers must accept slower generation speeds for the Plus tier.
