Over the last two decades, out-of-pocket (OOP) payments and external donations have contributed over 60% of current health expenditure (CHE) in low-income countries (LICs), in which OOP accounts for more than 40%, indicating a heavy reliance on individual contributions. Current studies on pricing models for healthcare largely focuses on achieving computational efficiency, but not the pricing effect on healthcare accessibility that influences healthcare improvement. Analysis of healthcare pricing models based on cost affordability can bridge this gap. Commonly used pricing models such as community rating (CR), dynamic pricing (DP), and OOP are considered. DP is a machine learning (ML) model based on Tanzania’s National Panel Survey (NPS) data, while the CR model is based on rates from the Tanzania’s Act Supplement for the mandatory public health insurance scheme of 2023, to the Ministry of Health (MoH). A pure premium approach for the DP model and current rates for the CR model are employed for comparative purposes. The results showed that CR does not significantly improve healthcare costs compared to DP ( \(\text{p-value = 0}\) ), conversely, DP significantly outperforms CR with a p-value \(3.49 \times 10^{ - 08}\) Moreover, the DP model remains superior to CR until a loading factor range of 5.7 and 6.4, where no significant difference, beyond which DP increases healthcare costs. Likewise, DP outperforms OOP until the loading factor range of 0.1 and 0.2, where costs are insignificantly different, above which DP increases costs. Load factor analysis confirms DP to significantly enhance healthcare accessibility by reducing cost compared to CR and OOP pricing models.