Markets
Desert Ant Labs Launches 18 On-Device Models Free to 100,000 Monthly Devices, With No Login and No Per-Token Price
Announced September 8 by Paul Veugen (previously the Detail video app), Desert Ant Labs shipped 12 stable and 6 beta audio, vision and text models behind unified SDKs: Voz transcription claimed 4.7x faster than Whisper, Clear at 302x realtime on an iPhone 16 Pro, Tongue at 0.933 language-ID accuracy in 2MB against 0.887 from a 293MB competitor, and Clips claimed 10x faster and 470x less energy than Sonnet. Redact catches 88.8% of personal data versus 91.1% for larger alternatives, which is the honest tradeoff being sold. The business model is the finding: free to 100,000 monthly active devices with no per-token metering, aimed straight at the Deepgram/AssemblyAI/Whisper-API line item.
Source
↳ Follow the thread