OpenAI has announced that it will be making it easier and cheaper for developers to insert large language models into apps and websites. Developers can now access ChatGPT and Whisper models through OpenAI’s API.
In a blog post, OpenAI said that it has achieved a 90% cost reduction for ChatGPT since December 2021 and that these savings will be passed on to API users. In addition, Whisper large-v2 model is now available in the API and is expected to produce faster and more cost-effective results.
The new ChatGPT API will cost $0.002 per 1,000 tokens, with the option to purchase the consumer-facing ChatGPT Plus for $20 monthly. Furthermore, OpenAI has plans to upgrade ChatGPT with continuous model improvements and the option to choose a dedicated capacity for more profound control over the models.
OpenAI also discussed its Whisper API, a hosted version of the open-source Whisper speech-to-text model launched in September. The Whisper API is an optimized version of the large model that offers faster and more convenient transcription. It will cost $0.006 per minute and will work in multiple languages, translating to English.
“We released a model, but that actually was not enough to cause the whole developer ecosystem to build around it,” said OpenAI President and co-founder Greg Brockman. “The Whisper API is the same large model that you can get open source, but we’ve optimized it to the extreme. It’s much, much faster and extremely convenient.”
Regarding the Whisper model, OpenAI stated in the blog, “Developers can now use our open-source Whisper large-v2 model in the API with much faster and cost-effective results.”
“We’re giving developers access to cutting-edge language (not just chat!) and speech-to-text capabilities,” said OpenAI in the blog post.
“ChatGPT API users can expect continuous model improvements and the option to choose dedicated capacity for deeper control over the models.”
In addition, developers who use the GPT-3.5-turbo model will always receive the recommended stable model while still having the flexibility to opt for a specific model version.
A new stable release of GPT-3.5-turbo will be updated in April.