# Artificial Analysis Speech Arena

Listeners hear the same sentence read by two AI voices without knowing which is which and pick the more natural one. The votes become a ranking of text-to-speech models.

Category: tts. Type: arena. Status: active.
Metric: Elo rating from blind listening votes (higher means preferred more often) (higher is better).
Source: Artificial Analysis — https://artificialanalysis.ai/text-to-speech/arena, retrieved 2026-11-17, license: proprietary (Terms of Use prohibit copying/redistribution).

Notes: Artificial Analysis Terms of Use (sec. 2.2(d)): 'except as expressly stated herein, no part of the Site may be copied, reproduced, distributed, republished...'; the Acceptable Use Policy also bans scraping. Their data platform page says external redistribution requires a commercial package. Rows left empty for that reason; see the live leaderboard at the source URL.
