[OpenAI Releases Two Speech-to-Text Models, Prices Drop by 25%] According to monitoring by Beating Insights, OpenAI has released two speech-to-text models: GPT Transcribe and GPT Live Transcribe. The former is designed for processing audio files and batch tasks, while the latter is tailored for real-time scenarios such as live captions and phone calls. Both models integrate audio themes, keywords, and language prompts, with a focus on improving recognition accuracy for short phrases, numbers, technical terms, diverse accents, and noisy environments. Artificial Analysis measured GPT Transcribe's word error rate at 3.31%, 0.7 percentage points lower than the previous generation GPT-4o Transcribe. Prices have also been reduced by 25%, with a charge of $4.5 per 1,000 minutes of audio. [Original Link]