[Feature Request] 请添加自定义模型的图片和文件输入功能

### 🥰 Feature Description

由于一些原因，例如使用openai格式的Gemini调用，在使用nextchat是只能使用自定义模型，而nextchat的自定义模型不支持图片或文件输入，这就限制了使用自定义模型时的可用性。

### 🧐 Proposed Solution

自定义模型时默认使能图片和文件输入，例如gpt-4o一样的界面即可。

### 📝 Additional Information

暂时没有

The issue names no files, tests, or entry points. Start by locating the custom-model configuration and the chat input UI, then trace how image and file attachments are handled for built-in models. Done means custom models expose the same image and file input controls as the GPT-4o interface.

Written by the indexing model from the issue text.