Open AI can be considered one of the leading AI-related companies in the world. They occasionally introduce new AI technology-related inventions. Social media users are also interested in these new technologies. However, some such social media posts are misleading. This is our investigation into them.
Social Media Posts
Social media users say Open AI’s new model O3 will change everything since artificial general intelligence (AGI) is finally here.
Users commented on it like below.
We decided to do a fact-check on this.
Fact Check
What is AGI?? AGI means artificial general intelligence. Artificial general intelligence (AGI) refers to the hypothetical intelligence of a machine that possesses the ability to understand or learn any intellectual task that a human being can. It is a type of artificial intelligence (AI) that aims to mimic the cognitive abilities of the human brain. More details can be reached here.
What is o3, and What Type of Invention Is it?
In December, Open AI conducted a 12-day program called 12 Days of Open AI. On the last day of the program, they introduced o3.
OpenAI has introduced its latest AI model, o3, which was created to improve cognitive capabilities like reasoning beyond its predecessor, o1. More details can be reached here. The o3 model is fluent in complicated works requiring higher cognitive skills, including STEM skills. It can do these tasks using a logical “private chain of thought” approach. This method allows the model to plan and think internally before generating responses, which paves the way for more reliable outputs.
Regarding ARC-AGI standards, o3 achieved three times the accuracy of o1 on this benchmark, which assesses an AI’s ability to handle novel and challenging mathematical and logical problems. More details can be read here. Early evaluations show that o3 scored a 2727 rating on Codeforces programming contests and 96.7 per cent on AIME 2024 mathematics problems. More information can be read here.
In December 2024, o3 and o3-mini underwent testing and were not publicly available. OpenAI opened it to the public on 31st January 2025. More details can be referred to here.
These advancements underscore OpenAI’s commitment to developing AI systems with improved reasoning abilities and safety measures, which will contribute to achieving artificial general intelligence (AGI).
However, experts caution that while o3 depicts significant milestones, AGI remains a distant objective, and ongoing research is essential to address the current limitations of AI reasoning models.
François Chollet, the creator of the ARC-AGI benchmark, expressed his views on this development. He warns against equating the o3 model’s performance with attaining AGI, stating, “Passing ARC-AGI does not equate to achieving AGI, and I don’t think o3 is AGI yet.” Chollet elaborates that o3 still fails on some straightforward tasks, indicating notable differences with human intelligence. More details can be read here.
Similarly, AI analyst Shelly Palmer, Professor of Advanced Media in Residence at Syracuse University’s S.I. Newhouse School of Public Communications and CEO of The Palmer Group, notes that although o3 represents a significant advancement in AI’s ability to tackle complex tasks, it does not equate to achieving AGI. Palmer suggests that while o3 and o3-mini demonstrate enhanced reasoning capabilities, they should not be misconstrued as evidence of AGI. Read more about this here.
Other AI experts, such as Don Lim, share similar comments. While acknowledging o3’s advancements, he emphasises that excelling in specific benchmarks does not necessarily mean an AI system possesses general intelligence. More details can be read here.
Conclusion
Despite o3 possessing significant advanced technology for cognitive thinking, it has not reached the levels of AGI or artificial general intelligence as per experts and other sources.
