[Feature]: Qwen3.5 support

[aropb]

Background & Description

Qwen3.5 support in llama.cpp already implemented:
ggml-org/llama.cpp#19468

API & Usage

No response

How to implement

No response

[Kago-Ryu]

Is it possible to support the new Qwen 3.5 models by recompiling the backend with the latest version of llama.cpp?

[martindevans]

Updating the backend always requires changes to the C# code, to adapt it to changes in the llama.cpp ABI. You can see an example of that in PR #1335 - that's updating to a version of llama.cpp from a couple of weeks ago. Once that's merged I'll probably look at updating again right away, since Qwen3.5 is so popular.

[Kago-Ryu]

Thank you!

[andrei-cb]

༼ つ ◕_◕ ༽つ

[martindevans]

#1356

This PR should add Qwen3.5 support. If you pull it and use it, please comment on the PR if it worked for you and what OS/hardware you used.

[jchristn]

+1 for needing this capability