Desert Ant Labs launched 18 on-device models (12 stable, six in beta) accessible via one SDK for Swift, Kotlin, and JavaScript. The lab says Voz transcribes 10 minutes of audio in two seconds on an iPhone, 4.7x faster than Whisper; Clear is a 9MB audio-enhancement model; Redact masks sensitive data in 27 languages; Tongue identifies 84 languages with a 2MB model. Full specs are on desertant.com and Hugging Face. Models are free up to 100k monthly active devices.
No score is assigned. Sources and their independence are shown in the citation chain below.