Apple is developing significant AI enhancements for iOS 26 and macOS Tahoe, including a high-performance transcription model. Tests show Apple's new transcription AI outpaces OpenAI's Whisper in speed and maintains comparable quality. In one test, transcribing a 34-minute podcast episode took only 45 seconds using Apple's model, while other AI transcription tools took nearly twice as long.
This new AI model will be available for developers to use within their applications. It is expected to improve the market for audio transcription by providing a robust and efficient solution. The transcription model isn't perfect, as it still has some difficulty with proper nouns.
Currently, these features are accessible to developers using the beta versions of iOS 26, macOS Tahoe and Xcode 26.