Try Voice Writer - speak your thoughts and let AI handle the grammar: https://voicewriter.io
Live updated leaderboard: https://voicewriter.io/speech-recognition-leaderboard
What are the best APIs for automatic speech recognition (ASR) in 2025? In this video, we benchmark all the major speech recognition APIs, including Google, Microsoft Azure, Amazon AWS Transcribe, startups Deepgram and AssemblyAI, the OpenAI Whisper model, and Google Gemini 1.5 Pro. We examine several different test conditions, including speech with noise, specialist vocabulary, and accents, and determine which APIs are best at handling each. Additionally, we evaluate real-time streaming and the generation of punctuation. Watch this video for an in-depth evaluation of which API to use for your project.
0:00 - Introduction
2:40 - Audio Data Selection
6:53 - APIs and Models
10:37 - Evaluation Metrics
12:06 - Main Results
16:40 - Real-time Streaming Results
21:48 - Final Winners