The economics of AI are starting to favor open models

Recent trends indicate that open-weight models like DeepSeek, Qwen, GLM, Kimi, and MiniMax are increasingly competitive against closed API alternatives, providing high intelligence at lower costs. This shift suggests that businesses may prioritize cost-effectiveness and control over marginal performance improvements, as the gap in performance between frontier models and strong open models narrows. For practitioners, this could mean a reevaluation of model selection criteria, emphasizing customization and privacy offered by open-source solutions.