OpenAI DevDay: OpenAI announced a wide range of AI products and improvements for the existing ones at its DevDay on November 6. One such AI service that deserves your attention is the new GPT-4 Turbo. According to OpenAI, the model is "more capable" yet "cheaper". For starters, the model has a massive content window of 1.28 lakh tokens.
So, how does this compare to the time-tested and powerful GPT-4? According to OpenAI, the seasoned model was announced in March and made available in July for developers. GPT-4 Turbo is the next generation of GPT-4, which is available in preview.
Internet Database
The new model is up-to-date with world events of up to April this year. On the other hand, GPT-4 sports a knowledge base of up to September 2021. This makes the newer upgrade more proficient with recent events.
Context Window And Cost
GPT-4 Turbo's context window can digest more than 300 pages of text in one prompt. In comparison, GPT-4 offers a 24-page (or 8,000 tokens) context window. A broader context window allows the model to generate more detailed responses after analysing the vast content library, compared to GPT-4.
"We also optimised its performance so we are able to offer GPT-4 Turbo at a 3x cheaper price for input tokens and a 2x cheaper price for output tokens compared to GPT-4," OpenAI said in its blog post. This will expand the accessibility of the model for more developers and users.
Release and Availability
GPT-4 Turbo can be accessed by paid developers by "passing 'gpt-4-1106-preview' in the API." The stable build is expected to be released in the coming weeks. GPT-4 is currently available for paying subscribers. On a related note, the AI research company also introduced GPT-3.5 Turbo with a default content window of 16,000.