OpenAI and Google LLC recently showcased their advanced reasoning models at the International Collegiate Programming Contest (ICPC), achieving gold-level performance. This prestigious coding competition features participants from around 3,000 universities. OpenAI’s model secured a perfect score, while Google’s Gemini 2.5 Deep Think solved 10 out of 12 problems. Held on September 4, the contest challenged 139 teams to tackle 12 complex problems within five hours. Notably, both AI solutions effectively solved a particularly difficult Problem C, which revolved around optimizing liquid distribution across reservoir networks. This milestone reflects significant advancements in abstract reasoning capabilities, essential not only in coding but also in various scientific and engineering fields. Additionally, OpenAI’s GPT-5 contributed by solving 11 problems with an experimental reasoning model tackling the final problem independently. These achievements highlight the growing role of AI in competitive programming and its implications for software development.
Source link
Share
Read more