This study employs the meta-analysis approach to assess the predictive accuracy of artificial intelligence (AI) models against traditional logistic regression in bankruptcy forecasting and to evaluate the contribution of market-based variables. It addresses two primary research questions derived from apparent contradictions in the extant empirical literature: (1) Which methodological approach—AI or logistic regression—provides superior predictive performance? (2) Does the incorporation of market variables enhance the accuracy of bankruptcy prediction models? A meta-analusis study identified relevant empirical studies published between 1990 and 2024. From this corpus, 573 effect sizes were extracted and analyzed using Comprehensive Meta-Analysis (CMA) software to test the corresponding hypotheses. The results indicate that AI-based models demonstrate statistically significant superiority in predictive accuracy compared to regression-based models. Furthermore, the analysis confirms that model performance is positively associated with the inclusion of a more number of market predictive variables of the bankruptcy prediction model. These findings clarify methodological efficacy and underscore the value of market data in bankruptcy prediction.