Kimi K2.5 is fourth place for intelligence right now. And it's not as good as the top frontier models at coding, but it's better than Claude 4.5 Sonnet. https://artificialanalysis.ai/models