Early results from the InnovationEval assessment indicate that current AI models are struggling to replicate a recent human algorithmic innovation. The findings highlight a persistent gap in artificial intelligence, demonstrating that these systems cannot yet match human-level algorithmic breakthroughs.