BYOK情况下支持更多的多模态模型；BYOK模式下上下文拓展（Support for more multimodal models under BYOK; context extension in BYOK mode）

Feature Requests
,
1

问题：

在BYOK模式下，许多多模态模型（如MiMo V2.5等）暂不支持多模态
并且在BYOK模式下，上下文只有180K

解决：

提供对于多模态大模型的多模态支持
提高BYOK模式下的大模型上下文限制

Problem:

In BYOK mode, many multimodal models (such as MiMo V2.5, etc.) do not currently support multimodal inputs, and the context length is limited to 180K.

Solution:

Provide multimodal support for large multimodal models and increase the context limit for large models in BYOK mode.

2

+1！！！上下文限制修改，支持一下应该很容易吧。我的deepseek api只能用180k上下文非常不爽啊。