Kokoro 82M v1.0, an open-source text-to-speech model, has been released and is considered the best downloadable option. It scores 180 ELO points lower than SpeechifyAI's Simba 3.2 in human preference evaluations. This performance is detailed on a media leaderboard, highlighting the current standing of open-source contributions in the text-to-speech field. AI
IMPACT This release provides a benchmark for open-source text-to-speech models, indicating areas where they lag behind commercial offerings.
RANK_REASON This is a release of an open-source model with performance benchmarks against a commercial competitor. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
