Our model (Simba 3.2) went #1 on the Artificial Analysis TTS leaderboard today and it's tied for #1 on realtime TTS on Voice arena!. That leaderboard is blind listening tests: thousands of people hear two voices, pick which sounds more human, no idea which model made which. As of today ours wins that more than anything from Google, ElevenLabs, Cartesia or anyone else!!
