OpenAI unveiled the GPT 4 Turbo, its most potent language model to date. The update was announced at OpenAI's first developer conference in San Francisco on 6 November.
The company founded by Sam Altman said that GPT-4 Turbo is an improved and more powerful version of the GPT-4 that was first released in March 2023.
With data going back to April 2023, GPT-4 Turbo—which is presently accessible through an API preview—has been trained, the company revealed on Monday at its inaugural developer conference. The GPT-4 version that was first released in March could only analyse data up to September 2021.
Although it did not provide a specific date, OpenAI intends to deliver a production-ready Turbo model in the next weeks, according to The Verge.
For developers, GPT-4 Turbo will also be more affordable to run.
With a 128K context window, which OpenAI claims is "equivalent to more than 300 pages of text in a single prompt," GPT-4 Turbo will also be able to 'see' more data. Large language models such as GPT are generally able to comprehend more of the query and provide more thoughtful responses within bigger context windows.
Furthermore, according to OpenAI, GPT-4 Turbo will be less expensive for developers to utilise than GPT-4, with input tokens costing three times as little as $0.01 and output tokens costing twice as much at $0.03.
In addition to enhancing the flagship model, OpenAI declared that it will offer enterprise users copyright indemnity via a programme dubbed Copyright Shield, emulating Google and Microsoft.
Copyright Shield will protect ChatGPT Enterprise's and OpenAI's developer platform's generally accessible functionalities.
