AI models demonstrate capability in functional circuit board design

Frontier artificial intelligence models have demonstrated the ability to design functional printed circuit boards (PCBs), marking a significant milestone in electronics engineering. Rather than using graphical interfaces, these models excel by working with declarative code representations, such as netlists and circuit description languages.

Performance is being measured via EEBench, a physics-graded evaluation framework developed by the team behind atopile. Unlike previous methods, EEBench uses SPICE simulations to verify that AI-designed circuits meet specific electrical requirements. Recent results show significant progress, with models like Claude Opus 5 and GPT-6 Astra scoring between approximately 60% and 70% on complex design tasks.

While these advancements allow for rapid prototyping and successful real-world applications, such as rechargeable LED components, experts note that human engineers remain essential. Current AI capabilities are viewed as powerful shortcuts for prototyping, but professional production runs still require the rigorous validation, edge-case handling, and independent testing provided by human engineers.