Overview
Qwen3-TTS-VC-Realtime is a snapshot version of Alibaba's real-time text-to-speech model for synthesis using replicated voices. It was released on November 27, 2025. The official release entry describes high-fidelity synthesis, adaptive tone based on the text, and handling of complex text. (help.aliyun.com)
Capabilities and uses
The model can synthesize speech in a voice created through a voice-enrollment service, making it suitable for applications that need real-time speech in a replicated voice. The release entry says the same voice can output 11 languages; a neighboring voice-enrollment entry says 10 languages, so the precise supported count should be verified against the applicable service documentation. (help.aliyun.com)
Limitations and availability
Voice replication requires a separate voice-enrollment step. Alibaba Cloud's international Model Studio retirement notice listed this exact model for retirement on October 10, 2026, subject to the actual change time, and recommended qwen3.5-omni-plus-realtime as a replacement. Check service availability and replacement capabilities before integration. (alibabacloud.com)
Pricing
- Alibaba Cloud international Model Studio, Singapore region: the retirement notice lists a historical input price of USD 0.13 per 10,000 characters for this exact snapshot. Output was not separately billed under the documented character-based TTS rule. This is a historical quotation, not a guarantee of current availability. (alibabacloud.com)
- The exact snapshot is no longer listed in the currently retrieved mainland-China pricing table; a current domestic RMB price could not be verified. (help.aliyun.com)
Pricing checked on November 17, 2026. The international and domestic sites are separate services and their prices should not be treated as interchangeable.