2024-09-13
TechCrunch
22 related
OpenAI claims that in a qualifying exam for the International Mathematics Olympiad, o1 correctly solved 83.3% of the problems, while GPT-4o solved only 13.4%
Sam Altman says it “doesn't constitute AGI” Poulami Saha / Financial Express : OpenAI makes big AI breakthrough, ChatGPT can now think and reason: Details Emilia David / VentureBeat : How to prompt on...
2024-07-19
CNBC
28 related
OpenAI launches GPT-4o mini, a smaller, cheaper offshoot of GPT-4o, replacing GPT-3.5 Turbo, with support for all multimodal inputs and outputs coming soon
each step more refined Aiming for the ‘perfect training set’ - the ultimate concentrate of all human knowledge and creativity Whether Gemini 1.5 Jeff Harris / @jeffintime : hidden gem: GPT-4o mini sup...
Loading articles...