On March 12, 2024, Bing and DuckDuckGo indexed a blog post from the ChatGPT maker OpenAI, and it suggests that the company is working on GPT-4.5 Turbo, an intermediary model that might come out before GPT-5 goes live. Whether this was an accident or intentional isn’t known yet. Nonetheless, the indexed post’s meta description reveals few critical details about the GPT-4.5 Turbo model.
GPT-4.5 Turbo Is Faster, More Accurate, And Scalable Than GPT-4 Turbo
Per a screenshot of the meta description circulating on Reddit, GPT 4.5-Turbo “surpasses GPT-4 Turbo in speed, accuracy, and scalability.” Further, GPT 4.5 Turbo can generate natural language or code with a 256K context window, higher than the 128K context window of the GPT-4 Turbo. However, other models, such as Gemini 1.5 Pro and Claude 3, already support an elaborate context window of 200K and one million tokens, respectively.
The Purported Model Will Come With Updated Knowledge Cutoff
Last but not least, the meta description also mentions how the purported language model will come with a knowledge cutoff of June 2024. This part of the text reveals that OpenAI intends to launch the GPT-4.5 Turbo model after June. Realistically, the model can come out in August this year, but the company has yet to confirm a launch date or timeline. Since the indexed blog post belongs to the company’s official website, the information leaked seems credible.
It is important to mention that the indexed page isn’t available anymore, indicating that it might have been an accident. Nonetheless, we now know that OpenAI plans to launch GPT4.5 Turbo before unveiling the next-generation GPT-5 model. Most recently, ChatGPT got the ability to remember things. OpenAI also released the Read Aloud feature for hands-free usage of the chatbot on the web and smartphones.
You can follow Smartprix on Twitter, Facebook, Instagram, and Google News. Visit smartprix.com for the most recent news, reviews, and tech guides.
