Today, March 14, marks a full year since OpenAI announced GPT-4, its advanced language model that has made incremental changes from its predecessor. At launch, the model demonstrated outstanding human-level results in various professional and academic tests and exams. Today, however, the company provoked speculation after it allegedly revealed details of the upcoming GPT-4.5 Turbo model for working with large languages in a blog post. The article, which is now unavailable, described this model as superior to its predecessor (GPT-4) in terms of speed, accuracy, and scalability.
People believe that the release is possible around July or August
The most interesting aspect of the leak is the mention of a “knowledge cutoff” in June 2024 for the model. This means that the artificial intelligence chatbot will have access to all information on the Internet before the June 2024 date. While some see this as a mistake, others believe that an OpenAI release around July or August is inevitable before the company moves to the next generation, GPT-5.
One of the big updates in GPT-4.5 Turbo is the 256k token context length window, which is twice the current 128k in GPT-4 Turbo. This move seems to be OpenAI’s response to competitors launching models with larger context windows. Ever since Google changed the name of Bard to Gemini and introduced new updates to its artificial intelligence model, the debate about whether Gemini should be considered a better LLM has not subsided. It seems that the new GPT model will level the playing field even for OpenAI, or even outperform Google’s AI chatbot.
OpenAI has not yet officially responded to the leak, leaving the GPT-4.5 Turbo status and June 2024 date shrouded in mystery. It couldn’t be a complete coincidence that the leak occurred on the first anniversary of GPT-4’s launch (could it?).
