OpenAI and Google DeepMind AI models win gold at ICPC

OpenAI and Google announced that their reasoning AI models won gold medals at the International Collegiate Programming Contest or ICPC, a prestigious coding competition held in Baku, Azerbaijan on September 4. While OpenAI said GPT-5 achieved a perfect score, Google’s Gemini 2.5 Deep Think solved 10 of the 12 problems in the contest. 

Google DeepMind hailed the victory in a blog saying, “Together, these breakthroughs in competitive programming and mathematical reasoning demonstrate Gemini’s profound leap in abstract problem-solving — marking a significant step on our path toward artificial general intelligence (AGI).” 

With close to 3,000 universities participating, just four out of 139 human teams won gold. 

Held every year, ICPC is known as the “coding Olympics,” and has a time limit of five hours to solve complex coding and algorithmic puzzles. 

Meanwhile, OpenAI used an experimental reasoning AI model that isn’t public yet along with GPT-5 in the contest. While GPT-5 solved 11 of the 12 problems, the last question which was also the toughest was answered by the experimental model. 

“We competed with an ensemble of general-purpose reasoning models; we did not train any model specifically for the ICPC,” Mostafa Rohaninejad, a technical staff at OpenAI said on X. 

Just two months ago, OpenAI and Google’s AI models had similarly won gold medals at the International Mathematical Olympiad.