[Inworld TTS-2 Officially Launched: Blind Voice Test Surpasses Sonic 3.6 to Take the Lead] Beating AI News Flash: Real-time voice AI company Inworld has officially launched Realtime TTS-2. In the latest Controlled Voice Arena by Artificial Analysis, it ranked first with an Elo score of 1123, surpassing Cartesia Sonic 3.6's score of 1119 by 4 points. This leaderboard tests different models by cloning the same set of 8 voices for blind testing, focusing on which sounds more natural and lifelike, including pronunciation, pauses, intonation, and rhythm. It incorporates real voice samples from earlier rounds as context, analyzing the user's tone, rhythm, and emotions to adjust how the next sentence is spoken. Developers can directly input natural language commands like "speak softly" or "sound tired but warm, as if just off work," and add sounds like laughter, sighs, and breathing. It also supports maintaining the same voice across languages and cloning voices using short recordings. Inworld's official statement claims this generation of models is designed specifically for real-time conversations, leveraging tone, rhythm, and emotional states in full voice interactions. [Original Link]