Researchers have introduced ASI-Bench, a new benchmark designed to evaluate the capabilities of artificial intelligence systems in exploring the unknown and conducting autonomous scientific research. The benchmark consists of 60 project-level tasks across 11 scientific domains, with progressively reduced human guidance. Results show that current AI systems remain heavily dependent on human guidance and are far from autonomously conducting end-to-end scientific research. ASI-Bench is open to the world, inviting researchers to contribute new tasks and challenge the limits of today's AI. --- Why it matters: This matters because it highlights the significant gap between current AI capabilities and the requirements for artificial superintelligence. Understanding this limitation can inform the development of more advanced AI systems that can truly explore and create new knowledge. Source: https://arxiv.org/abs/2608.17271