# Apple's New SpeechAnalyzer API Outperforms Whisper in Benchmark Tests Apple's latest speech recognition engine surpasses Whisper models in both accuracy and speed. - **Source:** [Hacker News: Front Page](https://brifing.si/source/b07aa772-b1a0-4a76-aa18-719c6f9e2dde) - **Published:** 2026-07-13 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/8871ff70-2329-428e-948c-4dae0531977b ## Apple's New Speech Recognition Engine Apple's new SpeechAnalyzer API marks a significant leap over its predecessor SFSpeechRecognizer. It achieved a 2.12% word error rate on clean speech in LibriSpeech tests, outperforming popular models like Whisper Small in both accuracy and speed. This demonstrates Apple's growing prowess in artificial intelligence technologies. ## Comparison with Whisper SpeechAnalyzer outperformed Whisper Small with a 2.12% error rate versus 3.74% on clean speech. It showed similar advantages in noisy environments and ran approximately three times faster. These results establish Apple's built-in engine as the top choice for Apple hardware users. ## Legacy API Performance Apple's old SFSpeechRecognizer API showed the weakest performance in tests, with a 9.02% word error rate on clean speech - even behind Whisper Tiny. The new API reduces this error rate by four times, providing compelling reasons for developers to migrate. ## Speed and Efficiency SpeechAnalyzer processed audio three times faster than Whisper Small while delivering more accurate results. All engines operated faster than real-time, processing one hour of audio in 1.5 to 5 minutes. Apple's optimizations show particularly impressive results on M2 Pro hardware. ## Language Support and Platform Dependence Whisper maintains advantages in supporting over 30 languages and being platform-agnostic. SpeechAnalyzer currently supports about 30 languages and is limited to iOS 26/macOS 26. However, for English on Apple devices, it has dethroned Whisper as the accuracy leader. ## Testing Methodology and Transparency Tests used the LibriSpeech corpus and showed consistency with OpenAI's published Whisper results. All raw transcripts were made publicly available for independent verification. The same code paths and normalization rules were applied across all engines for fair comparison. ## Implications for Developers Apple's SpeechAnalyzer API now represents the best choice for English speech recognition applications. While Whisper's multilingual support remains valuable, the time has come for Apple ecosystem developers to transition. This leap in accuracy and speed promises exciting developments for AI-powered audio processing. ## Sources covering this story - https://get-inscribe.com/blog/apple-speech-api-benchmark.html — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._