Google AI’s flu forecasting model was the top-performing model among 39 at forecasting flu-related hospital admissions during the 2025-26 season, the CDC found.
The CDC’s FluSight program combines weekly submissions from government, industry and academic teams that predict flu hospital admission for the current week and three weeks in the future. The agency uses the combined forecasts to show anticipated state-level demand for medical care. The end-of-season analysis was published Sept. 30.
It found that of 39 eligible models, Google’s AI, Google_SAI-FluEns, best matched the season’s observed hospital admissions.
The FluSight ensemble ranked seventh out of 39 models overall and was one of 12 models that consistently outperformed the baseline in all jurisdictions.
Of the 39 models, 33 performed better than the CDC’s baseline model.
