The auto model choice recently matched me with Claude Haiku.

The auto model choice recently matched me with Claude Haiku. Even with complex tasks, the reasoning en reponses of Haiku do not match the quality I expect. It also uses a lot of icons in it's response, making them hard to read. I allways need to switch to a medium teer model just to get another summary which is readable, because I'm to annoyed with the kinder garden icons.

---
| Field | Value |
| --- | --- |
| App version | 1.1.3 |
| OS | Windows 10.0.26100 |
| Theme | GitHub |
| Path | /chat |
| Tenure | Week 10 |

从 /chat 入口开始，在 Windows 10 上的应用版本 1.1.3 中重现自动模型选择，并将 Haiku 的响应与中等层级模型进行比较。Issue 未指明源文件或测试，并且必须在实现之前定义模型选择和响应格式的完成标准。

由索引模型根据 Issue 内容生成。