INESC MN participated in the international project Humanity’s Last Exam, an initiative by the Center for AI Safety designed to challenge the limits of today’s most advanced AI systems.

As part of the project, our team contributed with questions in the areas of electronics system integration, physics, and photonics, helping to test how far these technologies can really go. These questions were part of a larger effort by over 1,000 researchers and domain experts worldwide, aiming to probe the boundaries of AI capabilities with problems even the top models at the time—GPT-4o, Claude, Gemini, Mistral, and LLaMA—couldn’t fully solve.

This collaborative exercise highlighted not only the rapid progress of AI but also the human drive to explore, test boundaries, and continue pushing knowledge forward. At INESC MN, contributing to such an initiative was a unique opportunity to combine scientific curiosity with practical expertise in electronics, physics, and photonics.

👉 Learn more about Humanity’s Last Exam
👉 Full op-ed (in Portuguese) published in ECO