**[AI Voice Rankings See a New Leader: Cartesia Surpasses Qwen and ElevenLabs, Supporting 44 Languages]** According to monitoring by Dongcha Beating, Cartesia has released the next-generation real-time voice synthesis model, Sonic 3.6, which is now available in Beta. The model supports 44 languages and has secured the top spot in two TTS (Text-to-Speech) rankings by Artificial Analysis. In the Provider Voice test, which uses the native voices of each model for blind testing, Sonic 3.6 ranked ahead of Alibaba's Qwen-Audio-3.0-TTS-Plus, SpeechifyAI's Simba 3.2, and Google's Gemini 3.1 Flash TTS. In the Controlled Voice test, where all models use the same set of cloned voices to focus more on the models themselves, Sonic 3.6 still came out on top. Artificial Analysis measured Sonic 3.6's generation speed at 136.1 characters per second, compared to ElevenLabs' Eleven v3 at 46.7 and Gemini 3.1 Flash TTS at 24.2. The price is $49 per million characters, less than half of Eleven v3's cost, but still higher than Qwen's $27.6 and Simba 3.2's $10. [Original Link]