Prepare to be amazed by the newest and largest model from OpenAI, GPT-4.5! With improved writing capabilities, enhanced world knowledge, and a refined personality, this cutting-edge model is set to revolutionize the field of artificial intelligence. However, despite its impressive features, OpenAI warns that GPT-4.5 may not live up to expectations as a frontier AI model.
In a leaked document ahead of its official announcement, OpenAI clarified that while GPT-4.5 boasts 10 times better computational efficiency than its predecessor GPT-4, it does not introduce any significant new capabilities compared to previous reasoning releases like o1 or o3-mini. This raises the question – can GPT-4.5 truly be considered a frontier AI model?
Rumors have been circulating that OpenAI used their code-named Strawberry reasoning model (o1) to train GPT-4.5 with synthetic data. However, according to OpenAI’s statement on the matter, they employed new supervision techniques in combination with traditional methods such as supervised fine-tuning and reinforcement learning from human feedback (RLHF).
Despite some limitations in comparison to other models like ChatGPT Pro or Operator , leaks suggest that access to GPT-4.5 may be limited initially only for ChatGPT Pro users similar to how Operator was released previously . To learn more about the release plans for this groundbreaking model, tune into OpenAI’s upcoming stream.
As revealed last week in Notepad by yours truly , it seems we can expect not just one but two major launches from OpenAI – first with GPT-4.5 by end of February followed by their highly anticipated system integration project known as GPT-5 in late May. According to CEO Sam Altman’s previous statements, GPT-5 will incorporate OpenAI’s latest o3 reasoning model which was teased during their 12 days of Christmas announcements in December.
While o3-mini has already been released, the full-scale o3 model will only be available as part of GPT-5. This aligns with OpenAI’s ultimate goal to combine their large language models and pave the way towards creating an artificial general intelligence (AGI).
Stay tuned for more updates on this exciting development from OpenAI!