KT +SGLANG是否支持GGUF MOE模型读取，例如QWEN3.6 35B A3B GGUF 的 UNSLOTH量化

### Reminder

- [x] I have read the above rules and searched the existing issues.

### System Info

nothing

### Reproduction

```text
Put your message here.
```

### Others

_No response_

The issue names no files, tests, or entry points and provides no system information or reproduction. First clarify the expected support for the QWEN3.6 35B A3B GGUF model and its Unsloth quantization, then inspect the model-loading path and existing GGUF/MOE coverage; done should include a documented compatibility result or a defined implementation scope.

Written by the indexing model from the issue text.