biostudies-literatureUnknownLuebbert LNIGMS NIH HHSWe study large language models (LLMs) for front-line, pre-diagnostic infectious-disease triage, a critically understudied stage in clinical interventions, public health, and biothreat containment. We focus specifically on the operational decision of classifying symptomatic cases as <i>viral</i> vs. <i>non-viral</i> at first clinical contact, a critical decision point for resource allocation, quarantine strategy, and antibiotic use. We create a benchmark dataset of first-encounter cases in collaboration with multiple healthcare clinics in Nigeria, capturing high-risk viral presentations in low-resource settings with limited data. Our evaluations across frontier open-source LLMs reveal that (1) LLMs underperform standard tabular models and (2) case summaries and Retrieval Augmented GeneratiomedRxiv : the preprint server for health sciences2025.12.04.25341642https://www.ebi.ac.uk/biostudies/studies/S-EPMC12706609biostudies-literaturePandemic-Potential Viruses are a Blind Spot for Frontier Open-Source LLMs.PMC12706609T32 GM144273Nwachukwu NAchonduh-Atijegbe OOkocha YHenshaw ENwakpakpa EDada IEwah EFCronan PNosamiefan DOkoro SOlulaja OZhao KEktefaie YOkoeguale JAdebayo APOzonoff ARao ASOkokhere PLaning EWilkason Cvan Amburg CIraoyah KOyejide NHappi CBurris ALuebbert LSabeti PCSoumare HAmadi JOlayiwola FFry BfalsePandemic-Potential Viruses are a Blind Spot for Frontier Open-Source LLMs.We study large language models (LLMs) for front-line, pre-diagnostic infectious-disease triage, a critically understudied stage in clinical interventions, public health, and biothreat containment. We focus specifically on the operational decision of classifying symptomatic cases as <i>viral</i> vs. <i>non-viral</i> at first clinical contact, a critical decision point for resource allocation, quarantine strategy, and antibiotic use. We create a benchmark dataset of first-encounter cases in collaboration with multiple healthcare clinics in Nigeria, capturing high-risk viral presentations in low-resource settings with limited data. Our evaluations across frontier open-source LLMs reveal that (1) LLMs underperform standard tabular models and (2) case summaries and Retrieval Augmented Generatio2025-01-01T00:00:00Z2025 Dec2026-06-14T03:15:46.888Z2026-06-14T03:08:53.152ZS-EPMC127066094140967110.64898/2025.12.04.25341642