jasdooit

OpenAI’s GPT and Google’s Gemini, both artificial intelligence models, secured first and second place in a top-tier computer programming contest. After achieving a gold medal-level outcome at the International Mathematical Olympiad (IMO) in July, where they showcased their ability to solve mathematical problems, they have now demonstrated strong coding and programming skills.

On the 17th, according to local time, OpenAI shared on social media platform X that “our general-purpose reasoning model successfully addressed all 12 challenges from the 2025 International Collegiate Programming Contest (ICPC),” and mentioned, “This would be comparable to a top position achieved by a human.” The ICPC is the most renowned algorithm and programming competition globally, frequently called the “Programming Olympics.” Competitors are required to solve 12 problems within a five-hour timeframe.

GPT-5 from OpenAI addressed 11 issues, whereas an unlaunched reasoning model managed the single problem that GPT-5 couldn’t handle. OpenAI mentioned, “A year ago, our reasoning models were considerably less intelligent and much more costly. However, we now have GPT-5—a model that is more intelligent, quicker, and more economical—allowing advanced AI reasoning.”

Google’s “Gemini 2.5 DeepSync” successfully addressed 10 of the 12 challenges, securing a second-place performance relative to human teams. Google DeepMind stated, “In addition to math and programming, this accomplishment highlights a significant new ability in abstract thinking,” and noted, “The skill to grasp complicated issues, create multi-step logical strategies, and execute them perfectly is essential in numerous scientific and engineering areas, like the development of new medicines or semiconductor design.”

Previously, both companies’ AI models showcased mathematical problem-solving abilities at the IMO by accurately solving five out of six problems. Both the questions and responses were presented in natural language format, and the models developed solution processes similar to those used by humans. A representative from the technology sector stated, “The fact that AI performs well in elite competitions suggests that its creative reasoning or problem-solving skills have attained the highest level.”