Hugging Face has unveiled the Open TTS Leaderboard, a new evaluation framework designed to standardize and accelerate assessment of text-to-speech models. The platform addresses a critical gap in the rapidly expanding TTS field: while the Hugging Face Hub now hosts more than 8,000 TTS models, evaluation remains fragmented and unstandardized. Existing arena-based leaderboards like TTS Arena v2 and Voice Arena rely on human preference voting to rank models, a process that can take weeks and disproportionately favors commercial models with dedicated API infrastructure over open-source alternatives. The Open TTS Leaderboard leverages objective metrics to evaluate models across multiple dimensions: intelligibility through word and character error rates (WER/CER), inference speed via real-time factor (RTFx) and time-to-first-audio (TTFA), and speaker similarity using WavLM embeddings. This metric-driven approach dramatically accelerates evaluation from weeks to hours, using an H200 GPU for benchmarking. The leaderboard supports multilingual evaluation across multiple languages and includes dedicated voice cloning comparisons, enabling more comprehensive assessment of model capabilities. While the Hugging Face team acknowledges that objective metrics don't directly measure naturalness or expressiveness—and thus don't replace human preference rankings—they position the leaderboard as a complementary tool to inform both researchers and existing voting-based leaderboards. The platform features a "Listen" tab allowing users to compare model outputs directly and provide community feedback that may be incorporated into future rankings.