We have some exciting news and updates from OpenAI to share with you today. There have been significant advancements and improvements in the API world, and we're here to break it down for you.
OpenAI has officially made their most advanced language model, GPT-4, generally available to all existing API developers with a record of successful payments. The GPT-4 API, which comes with an 8K context, is expected to be available for new developers by the end of the month. With its wide range of capabilities, GPT-4 is set to revolutionize the development of innovative products.
In addition to GPT-4, OpenAI has also made the GPT-3.5 Turbo, DALL·E, and Whisper APIs generally available, indicating their readiness for production-scale use. The fine-tuning feature for these models is currently being worked on, and it is expected to be available later in the year.
OpenAI is encouraging a transition from the freeform text prompt of the Completions API to the more structured interface of the Chat Completions API, which now accounts for 97% of OpenAI's GPT usage. The Chat Completions API offers a more powerful and flexible interface, allowing developers to build more diverse conversational experiences and completion tasks. OpenAI is investing more resources in this area to further enhance the developer experience.
As part of this shift, OpenAI has announced a deprecation plan for older models in the Completions API. From January 4, 2024, older completion models will no longer be available and will be replaced by newer models. Applications using the stable model names for base GPT-3 models (ada, babbage, curie, davinci) will be automatically upgraded. Developers using other older completion models will need to manually upgrade their integration by the specified date.
If you have been using fine-tuned models, you'll need to fine-tune replacements atop the new base GPT-3 models or the newer GPT-3.5 Turbo or GPT-4 models. OpenAI has promised to provide support during this transition.
Users of older embeddings models will need to migrate to the text-embedding-ada-002 model by January 4, 2024. This model, released in December 2022, has proved more capable and cost-effective than previous versions.
The Edits API and its associated models will need to be migrated to GPT-3.5 Turbo by January 4, 2024. Feedback from the Edits API was instrumental in the development of gpt-3.5-turbo and the Chat Completions API.
It's always a thrilling ride to keep up with the evolving world of AI. Here at DevPicker, we'll continue to keep you updated with the latest news and changes in the AI landscape. Remember, the only constant in technology is change!
Happy developing!