OpenAI has unveiled GPT-4 Turbo, the latest iteration of its cutting-edge language model, which offers both enhanced capabilities and cost savings. This advanced model has been trained on data updated until April 2023, surpassing its predecessor, GPT-4, which had knowledge only up to September 2021.
We're rolling out new features and improvements that developers have been asking for:
— OpenAI (@OpenAI) November 6, 2023
1. Our new model GPT-4 Turbo supports 128K context and has fresher knowledge than GPT-4. Its input and output tokens are respectively 3× and 2× less expensive than GPT-4. It’s available now to…
GPT-4 Turbo introduces a context window of up to 128,000K, equivalent to roughly 300 pages of text, providing deeper comprehension and more detailed responses to user queries. Beyond text-based inputs, the model can process image prompts, generate spoken responses, and even assist with coding tasks in various programming languages.
Notably, GPT-4 Turbo is available as a preview API for developers at a significantly reduced cost, with input tokens priced at just $0.01 per 1,000 tokens and output tokens at $0.03 per 1,000 tokens. This pricing makes GPT-4 Turbo three times more economical compared to its predecessors.
Bradley Rhodes
Furthermore, OpenAI is extending improved capabilities to GPT-3.5 Turbo users, enabling them to enjoy a 16,000K context window and new features similar to those of GPT-4 Turbo.
OpenAI's commitment to innovation also includes providing copyright indemnity to customers, protecting them against potential legal claims related to copyright infringement. This initiative, called Copyright Shield, aligns OpenAI with industry leaders like Google and Microsoft in safeguarding customers from such legal challenges.