OpenAI has announced the general availability of GPT-4, their latest language model, to paying API customers. GPT-4 is hailed as OpenAI’s most capable language model to date, and since its release in March, it has powered various applications and services, including visual accessibility, wealth management, language preservation, and language learning. In a blog post, OpenAI provides insights into the availability of GPT-4 and outlines the deprecation of certain APIs by January of the following year.
OpenAI states that existing API developers with a payment history can access GPT-4 with 8K context. The context window determines the amount of text GPT-4 uses to generate output, and the 8K context option corresponds to approximately 13 pages of text. OpenAI plans to raise rate limits and extend access to GPT-4 for new developers by the end of July, based on compute availability. Additionally, OpenAI intends to make GPT-3.5 Turbo, DALL-E, and Whisper APIs generally available to developers.
Also Read: OpenAI unveils “Incognito-Mode” for ChatGPT
Deprecation of GPT-3 and Variations
Starting from January 4, 2024, GPT-3 and its variations using the Completions API will be retired to free up computing capacity. Although the API will still be accessible, it will receive a legacy label and be replaced by a new model. Applications using stable names for base GPT-3 models (such as ada, babbage, curie, and davinci) will automatically be replaced by their successors (ada-002, babbage-002, curie-002, and davinci-002).
Deprecation of Embedding API
By January 4, 2024, developers utilizing older embedding models will be required to migrate to the new embedding model introduced by OpenAI in December 2022. OpenAI asserts that the new model is more capable and cost-effective than its predecessors. While the new model already accounts for 99.9% of embedding API usage, OpenAI acknowledges that switching models can be costly and inconvenient for some developers. Consequently, OpenAI commits to covering any costs associated with re-embedding content.
Conclusion: Advancements in Language Processing with GPT-4
With the general availability of GPT-4, OpenAI has made significant strides in advancing language processing capabilities. API customers can now harness the power of GPT-4 for a wide range of applications. As GPT-3 and older embedding models are phased out, developers are encouraged to transition to the new models to benefit from improved performance and cost-effectiveness. OpenAI’s commitment to supporting developers during this transition highlights their dedication to empowering innovation in natural language processing.