Construction & Engineering
AI Models Score as Low as 2% on Construction Takeoff Test
A new benchmark found that general-purpose AI models performed poorly on construction takeoff tasks involving technical drawings and measurements.
Togal.AI reported that general-purpose AI models scored between 2% and 34% on its Construction’s Last Exam benchmark for construction takeoffs. The benchmark contains 33 tasks requiring models to interpret drawing sets, identify construction elements and produce measurements used in estimating. Models were given raw PDF drawings and could use built-in vision and coding capabilities, but not third-party tools or agentic workflows. The results suggest that general-purpose systems still struggle with technical plan interpretation, an area where construction estimators require high precision. Togal.AI said its own machine-vision models were trained on tens of millions of construction plans, underscoring the importance of industry-specific data and workflow design when applying AI to estimating.
Original source: Construction Briefing