PCB-Bench: Benchmarking LLMs for PCB Placement and Routing (ICLR 2026)
PCB-Bench is a benchmark designed to evaluate large language models on printed circuit board placement and routing tasks, accepted at ICLR 2026. It provides a standardized dataset and metrics for assessing LLM performance in electronic design automation.
Background
- PCB-Bench is a benchmark (accepted at ICLR 2026) that tests LLMs on printed circuit board (PCB) placement and routing — arranging components and drawing electrical connections on a board.
- Most LLM benchmarks focus on text, math, or code; PCB design involves spatial reasoning and real-world constraints (electrical, thermal, manufacturability), making it a novel and harder test of AI capability.
- PCB layout has traditionally been a manual, expert-driven task, only partially automated by algorithmic tools. This benchmark asks whether LLMs can handle that design challenge.
- The repo (from DigiLab) provides PCB design files, prompts, and scoring scripts so researchers can evaluate different models on this engineering task.