Alibaba's Qwen-Audio-3.1-TTS model has achieved the top global ranking on the Artificial Analysis Controlled Voice Arena leaderboard. This new generation text-to-speech model supports multilingual synthesis and allows for natural voice migration across languages, focusing on contextual appropriateness and emotional expression. The Qwen-Audio-3.1 series, including ASR and real-time interaction models, is now available via API on the Qwen AI platform, with several related open-source models also gaining significant traction on GitHub. AI
IMPACT Sets a new benchmark for multilingual and context-aware speech synthesis, potentially influencing future model development and applications.
RANK_REASON Model achieves top ranking on a specific AI benchmark leaderboard. [lever_c_demoted from research: ic=1 ai=1.0]
- Alibaba Group
- Artificial Analysis
- Controlled Voice Arena
- CosyVoice
- GitHub
- Qwen AI platform
- Qwen-Audio-3.1-TTS
- Qwen-Audio-Agent
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →