New Delhi. Ever since the chatgpt of OpenAI has come into existence, he remains engaged in the news since then. Now the new version of the Chat GPT has given a strong marks in the country’s most difficult exam and gave the discussion to the discussion. Openai’s latest AI model “Chatgpt O3” took the JEE Advanced 2025 exam and in this she got almost full marks. Joint Entrance Examination (JEE) Advanced is the most competitive university entrance examination in India, which is organized for admission in IITs.
Ashwi rebuilt the actual test conditions for AI. In his blog “Helper”, he told the details that the model was motivated to work “like a JEE candidate” and every question was asked to resolve individually and without the help of web search or external python tools. To eliminate any memory bias, every question was presented in a new chat session and no improvements or indications were given during the process.
Despite these stringent sanctions, Chatgpt O3 showed remarkable efficiency. AI scored 60 marks in Chemistry and Mathematics in the second phase of the simulated examination, while only a few marks in Physics and earlier sections. This unprecedented performance of an AI chatbot in this high-day, human-centric examination highlights the rapidly developing abilities of Artificial Intelligence and inspire important discussions about its possible impact on the definition of education, competitive evaluation and “intelligence”.
In a new published research paper “The Ilusian of Thinking”, Apple’s team argues that even today’s most advanced language models are not engaged in as much true logic as it is commonly considered. Their study suggests that, although these models can effectively simulate intelligence, their capabilities break to a great extent when facing deep and complex challenges.
