OpenAI has been in the news ever since it launched its latest advanced large language model GPT-4, last month. While a few users are still trying to catch a glimpse at the latest GPT-4, the rumors for the next LLM, GPT-5 have already started to emerge in the market. The rumors say OpenAI is expecting to complete the AGI training of GPT-5 by the end of 2023.
Currently, there haven’t been any comments surpassed by OpenAI regarding this project. The rumors of GPT-5 started after a developer Siqi Chen suggested in his tweet suggesting a few individuals in the company feel GPT-5 is capable of achieving AGO (Artificial general intelligence) a milestone that indicates the AI is capable of displaying human-like intelligence and understanding.
- OpenAI is rumored to launch a new large language model, GTP-5 by the end of 2023, according to this report.
- There have been rumors regarding the possibility of GPT-5 attaining AGI (Artificial general intelligence) meaning obtaining the same understanding and intellect as a human.
- OpenAI has not stated any confirmation regarding the rumors of GPT-5 and has not confirmed whether GPT-5 is at work or not.
Will GPT-5 launch by the end of 2023?
Last month, OpenAI released its advanced language model GPT-4, with an introduction to its latest feature “Input vision” along with capabilities of displaying human-like performance on a variety of academic and skilled benchmarks. Now, there have been rumors about launching GPT-5 by the end of 2023.
The rumors of GPT-5 started after a report by BGR-based tweets was released by developer Siqi Chen, stating OpenAI should be able to complete the training period of its next large language model GPT-5 by the end of the year.
As of now, OpenAI has not stated any information regarding its next large language model, GPT-5, so currently, there is no data available regarding the new features, changes, or improvements. Siqi Chen tweeted on Twitter saying, “OpenAI expects to achieve AGI.”
AGI stands for Artificial general intelligence. So, if GPT-5 clears or reaches AGI, it indicates the large language model has accomplished human understanding and intelligence.
OpenAI’s GPT-4 is more advanced and more capable than its previous model GPT-3.5, which was powered by the popular AI chatbot, ChatGPT. GPT-4 has introduced new features such as the ability to analyze images along with texts, perform complex tasks, be creative, and be more accurate than its predecessor.
We can assume GPT-5 will be more advanced than its predecessor and would mark a leap forward, although there is no information regarding its features, improvements, or changes available for it’s unclear how that leap will move forward.
Later Siqi Chen backtracked a bit on his latest tweet by saying, “To be precise, I didn’t mean reaching AGI with GPT-5 is a consensus belief within OpenAI, but non-zero people there would think it will get there.”
“Non-zero people” consider GPT-5 might attain AGI is quite different from stating “OpenAI expects it to achieve AGI.”
Reports of GPT-5 have yielded conflicting angles. In addition, TechRadar shared a piece earlier this week titled, “ChatGPT has passed the Turing test and if you’re freaking out, you are not alone.” Later there was an article published by Tom’s Guide with the title, “No, ChatGPT didn’t pass the Turing test – but here’s when it potentially does it.”
Since the rumors of GPT-5 aren’t confirmed by its creator OpenAI, it’s best not to raise any expectations until it’s confirmed by the official source itself. If GPT-5 is launched by the end of 2023, then it will be way more capable and advanced than its previous model GPT-4. Although any discussion about AI obtaining human-level understanding and intellect might have to wait for now.