Newest papers

ProteinGeoBench: Frontier Structural-Biology Agents Are Fluent at Lookup but Lost in Geometry

Matt Turk, Brandon L. Neel, Diana M. Lopez, Zachary D. Smith, Grace N. Johnecheck, James B. Golden · 2026 · arXiv

DOI: 10.48550/arXiv.2610.26829 · open-access PDF

French

Les agents frontaliers récupèrent bien les métadonnées structurelles et calculent la géométrie des coordonnées de manière moins fiable. Nous demandons combien d'un score de référence reflète ce fossé et combien reflète la configuration d'évaluation autour du modèle. Nous étudions ProteinGeoBench, une suite de 131 tâches de structure de protéines, en utilisant 4,249 déploiements de quatre modèles frontaliers et 4,486 essais supplémentaires sur 26 modèles. Trois des quatre modèles frontaliers tombent dans la limite de 0,031 l'un de l'autre (0,785 à 0,816) sans aucune différence entre eux résolue, tandis que le quatrième note 0,707 et les trois lacunes à elle excluent zéro.

Machine translation (Meta M2M100); the English original follows.

Abstract

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 four frontier models fall within 0.031 of one another (0.785 to 0.816) with no difference between them resolved, while the fourth scores 0.707 and all three gaps to it exclude zero. Three of the four do b

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.26829, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange