ProteinGeoBench: Frontier Structural-Biology Agents Are Fluent at Lookup but Lost in Geometry
Frontier agents retrieve structural metadata well and compute geometry from coordinates less reliably. We ask how much of a benchmark score reflects that gap and how much reflects the evaluation configuration around the model. We study ProteinGeoBench, a suite of 131 protein-structure tasks, using 4,249 rollouts from four frontier models and 4,486 further trials across 26 models. Three of the…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.