调用记录 把 WorkBuddy 后台这 126 次调用的积分消耗数据拆分来看，各个模型的用量占比会更直观：其中 glm-5.2 和 minimax-m3 的积分消耗最高。在 auto 自动模式下，hy3 会自动识别任务并分配对应模型，整个调用链路基本是固定的。也就是说，即使提前给 WorkBuddy 设置“遇到某类任务时提供切换模型选项”，实际运行中也无法真正生效。对于用户来说，要么不选择 auto 自动分配模式，要么就只能接受 auto 的默认模型分配结果。:false,"href":"","boxShadow":"","width":509,"aspectRatio":"0.671254","status":"success","showText":true,"isPercentage":false,"percentage":0,"isHoverDragHandle":false}},{"type":"image","attrs":{"id":"5df1472d-d01e-45d5-87ed-791df66bfe4e","src":"https://developer.qcloudimg.com/http-sa ve/audit-11165492/c53ca08a0014214901955bb94496925b.png","extension":"png","align":"center","alt":"","showAlt":false,"href":"","boxShadow":"","width":1018,"aspectRatio":"1.985885","status":"success","showText":true,"isPercentage":false,"percentage":0,"isHoverDragHandle":false}},{"type":"paragraph","attrs":{"id":"e572cf84-646a-414e-9baa-93504a66a303","textAlign":"inherit","indent":0,"color":null,"background":null,"isHoverDragHandle":false},"content":[{"type":"text","text":"workbuddy工作的链路，"}]},{"type":"image","attr图片 总结一下：目前积分消耗规则与最终生成质量之间的对应关系并不透明，也没有被充分说明；而从实际使用体验来看，更关键也更现实的问题在于，WorkBuddy 积分本身确实远远不够用，尤其在高频调用和自动模型分配场景下更为明显。