one year on
OpenAI unveils GPT-4 Turbo with 128k context, cuts prices at first DevDay
The company also announced an experimental program to fine-tune GPT-4 and doubled the tokens-per-minute rate limit for paying GPT-4 customers.
At its first-ever developer conference today, OpenAI launched GPT-4 Turbo, a faster and cheaper version of its flagship model with a 128,000-token context window — enough to digest the entire text of “Harry Potter and the Prisoner of Azkaban” in one go. The company says the model costs $0.01 per 1,000 input tokens, roughly three times cheaper than GPT-4, and has knowledge up to April 2023.
OpenAI also launched an experimental access program for fine-tuning GPT-4 and doubled the tokens-per-minute rate limit for paying GPT-4 customers.
The record
One year later — open only if you can handle spoilers
Eleven days later, the OpenAI board fired Sam Altman, triggering a chaotic weekend that nearly saw him move to Microsoft. He was reinstated five days later, and the GPT Store eventually launched in January 2024 with less fanfare than the revenue-share promises suggested.