This paper introduces SAFARI-P (Swahili-Focused Adaptive Framework for Accelerated Reinforcement in Intelligent Python Education), an innovative system integrating Generative Adversarial Networks (GANs) and Confident Learning (CL) for Python programming education in Swahili. The framework comprises three key components: an Adversarial Code Generation System (ACGS) for creating culturally relevant code snippets, a Confident Learning-based Assessment Module (CLAM) for nuanced code evaluation, and a Cultural Context Integration Engine (CCIE) for seamless incorporation of local cultural elements. In a 16-week study involving 500 students across Kenya, Tanzania, and Uganda, SAFARI-P demonstrated significant improvements over traditional methods: a 27% increase in Python proficiency (p < 0.001, d = 2.73), a 32.6% improvement in problem-solving efficiency (p < 0.001, d = 2.58), Through a mixed-methods approach, we illustrate how SAFARI-P’s adaptive learning paths, culturally relevant content generation, and real-time feedback mechanisms directly contributed to these improvements, making programming education more accessible and effective for Swahili-speaking learners.