Artificial Analysis adds text-to-speech leaderboards for nine languages

Artificial Analysis has extended its Controlled Voice Arena with separate text-to-speech leaderboards for Japanese, Mandarin Chinese, Hindi, Spanish, German, French, Portuguese, Vietnamese, and Arabic. Models are compared using standardized cloned voices and prompts written natively in each language. First-language speakers cast the preference votes, and Elo scores are calculated independently for each language. The nine boards include more than 100,000 votes and rank 15 to 24 public models per language.

Cartesia’s Sonic family leads eight of the nine leaderboards. Sonic 3.6 ranks first in seven languages, and Sonic 3.5 leads Portuguese. Inworld AI’s Realtime TTS-2 ranks first in Mandarin at 1,185 Elo, ahead of Sonic 3.6 at 1,146 and StepFun’s StepAudio 2.5 TTS at 1,130. Realtime TTS-2 also ranks first in English (US). ElevenLabs’ Eleven v3 family ranks in the top three in seven of the nine languages through Eleven v3 and Eleven v3 Conversational.