Agent Arena reveals top AI models for complex tasks as Google and Xiaomi gain ground

Google and Xiaomi gained ground in Agent Arena’s ranking of AI models for complex tasks, a shift that could push the industry toward balancing performance with affordability. The article does not disclose specific token prices, total value locked, trading volumes, market capitalization or funding figures, nor does it name individual protocols, chains or tokens beyond the model developers. The result matters because cheaper models that handle demanding workloads could pressure pricing across AI services and change how developers choose models. What to watch next is whether Google and Xiaomi sustain their gains in subsequent Arena updates and whether cost becomes a larger factor in model selection.