Background/purpose. Innovative thinking and meta-skills are needed to solve difficult problems in the 21st century, but standard training techniques often fail. The fact that only 6% of pupils are innovative thinkers highlights a major educational gap. The GENICS learning approach (Grouping, Exploring, Discussion, Individual Activity, Combining, Sharing) is contrasted to Problem-Based Learning (PBL) in tailored instruction to improve these skills. Materials/methods. A quasi-experimental approach with pretest-posttest control groups was used with 432 Grade X students from SMA Negeri 5 Jambi City. The experimental group (n=36) used GENICS while the control group (n=36) used PBL. Validated essay tests and observation sheets assessed creative thinking (fluency, flexibility, originality, elaboration) and meta-skills (self-management, social intelligence, invention). Methods included descriptive statistics, MANCOVA, and effect size estimations. Results. The GENICS learning model improved creative thinking by 93.48% and meta-skills by 26.78%, outperforming PBL by 56.42% and 4.70%, respectively. Elaboration (153.49%), originality (116.77%), critical thinking (44.37%), and empathy (35.26%) improved. GENICS showed superior efficacy with a medium effect size ([eta-squared] = 0.097, p = 0.036) in multivariate tests. Conclusion. The GENICS model's structured yet flexible stages use differentiated learning ideas that exceed PBL for cognitive and noncognitive skills. Complete 21st-century skill development is provided by the tiered system, especially the Combining and Sharing stages that encourage collaborative elaboration. Findings suggest curriculum designers should use differentiated models. Single-school and short interventions are drawbacks. Examine GENICS' effects on emotional intelligence, retention, and scalability in education. Study emphasizes systematic instructional design for mixed classes.