The rapid advancement of artificial intelligence (AI) is ushering in a new era for technology, progressively rendering traditional calculators obsolete. Modern AI models are now capable of solving even the most complex mathematical problems, accelerating developments across mathematics, science, and computer coding.

Just a few years ago, calculators were the primary tool for mathematical solutions, but AI has now taken their place. AI models developed by Google, OpenAI, and DeepSeek have achieved a significant milestone by scoring at gold-medal levels in the International Mathematical Olympiads. DeepSeek's Math-V2 AI model, developed in China, demonstrated success in the competition by not only providing correct answers but also transparently explaining the underlying logic—a standard achieved by only 8 percent of human participants. However, unlike the five human participants who achieved top-tier success in the competition, no AI model secured a perfect score.

Expert Perspective

Gregor Dolinar, President of the Olympiad and an engineering professor at the University of Ljubljana in Slovenia, noted the gradual disappearance of physical calculators. Describing AI's progress in this field as “fascinating,” Dolinar told AFP news agency, “You needed scientific calculators in the past, but now it’s easier to ask AI.” He added, “If you ask the question correctly, AI can solve abstract, logical questions and show how it reached the result.”

When OpenAI launched the AI boom with ChatGPT in 2022, chatbots could answer questions, write poetry, summarize articles, and even generate small computer codes. However, they frequently struggled with mathematics. Over the past two years, companies like Google and OpenAI have developed systems more adept at tackling complex mathematical problems that are beyond the average human's ability. Last year, Google DeepMind introduced math-specific systems like AlphaGeometry and AlphaProof. These are not chatbots, requiring mathematicians to translate problems into Lean, a computer programming language designed for mathematical problem-solving, before they can be addressed.

Self-Correction and Future Potential

OpenAI has stated that its latest AI model, GPT-5, is deepening research in mathematics, biology, and physics. On November 20, the model assisted a mathematician at Columbia University in solving a previously unsolved problem. DeepSeek is now focusing on providing question answers through a “self-correction” model. According to the Chinese company, this method is designed to function even when no pre-existing solutions for problems are available. This self-correction capability allows AI to evaluate the consistency and validity of its own reasoning, ensuring accuracy for known solutions and reliability when tackling new or unsolved mathematical challenges. Researchers suggest these results indicate a promising process where self-verifying mathematical reasoning could pave the way for even more capable AI systems.

News about Technology