Nvidia and ChipAgents release Build-A-Chip-Bench for chip design agents
Build-A-Chip-Bench tests whether AI agents can autonomously design chips; the best agent scored 16.6%, with about 40% of tasks passing functional verification.
🚨BREAKING: Nvidia and ChipAgents have launched Build-A-Chip-Bench, a new benchmark testing whether AI agents can autonomously design entire chips. Even the best-performing AI agent, powered by Anthropic's Claude Fable 5.1, scored just 16.6%. Across tested agents, roughly 40% of tasks passed functional verification, but only 10% met the physical requirements needed for a viable chip design. ChipAgents CEO William Wang says AI has already cut chip development timelines from at least two years to around one year. He expects further advances to bring that down to just four to six months within the next year.