The ChatGPT and Whisper APIs have been released to allow developers to integrate them into their applications. These APIs enable apps to use features such as natural language processing (NLP) and text-to-speech (TTS).
The API of ChatGPT provides a conversational interface capable of understanding and responding to natural language queries. Whisper’s API includes voice synthesis, which can convert text into natural-sounding speech. These APIs could be used in a variety of industries, including customer service, education, and entertainment.
OpenAI’s new ChatGPT API model is known as “gpt-3.5-turbo,” and it replaces its previous “best” LLM API, “text-davinci-003.” It costs $0.002 per 1,000 tokens (roughly 750 words), which OpenAI claims is about ten times less than its existing GPT-3.5 models. “We’ve achieved 90% cost reduction for ChatGPT through a series of system-wide optimizations since December,” writes OpenAI on its API announcement page.
The Whisper API, on the other hand, is available for $0.006 per minute and is based on the open source whisper-large-v2 model. It can convert audio to text or transcribe at a rate comparable to a skilled human transcriptionist, even with difficult audio, and accepts inputs in M4A, MP3, MP4, MPEG, MPGA, WAV, and WEBM formats.
Early ChatGPT API users include Snapchat with its “My AI” bot, Quizlet, an educational platform that reportedly helps students study, and Instacart, which plans to add “Ask Instacart” later this year to allow customers to ask about food.
The sources for this piece include an article in ArsTechnica.
