Atlas

Benchmarks

← All benchmarks

BioSecBench-Surveillance

Science · 2026-07-09

Pathogen genomic surveillance benchmark on real sequencing data, spanning variant detection, taxonomic classification, antimicrobial-resistance characterization, source tracking, and anomaly detection across wastewater, clinical, and environmental samples.

Top models (higher is better)

ModelScore
Grok 4.553.3
GPT-5.6 Sol52.5
GPT-5.6 Terra51.0
GPT-5.6 Luna50.3
Opus 4.850.2
GPT-5.550.2
Sonnet 549.8
Opus 4.749.6
Sonnet 4.648.6
Gemini 3.5 Flash47.1
Opus 4.646.7
Gemini 3.1 Pro Preview45.3
GPT-5.441.3
Grok 4.316.0
Grok 4.2013.7
Loading Atlas data…