Update july 2025

[martindevans]

Updated llama.cpp binaries to 11dd5a44eb180e1d69fac24d3852b5222d66fb7f, compiled with this run.

Testing:

• Windows CPU
• Windows CUDA
• Windows Vulkan
• Linux CPU (x64)
• Linux CPU (ARM64)
• Linux CPU (musl)
• Linux CUDA
• Linux Vulkan
• MacOS

[m0nsky]

Unit tests passed on Windows CUDA & Linux CUDA.

Test application is running fine on:

• Windows CPU
• Windows CUDA
• Windows Vulkan
• Linux CPU
• Linux CUDA
• Linux Vulkan

[jihadkhawaja]

Any possibility to include llama win-arm64-cpu binary in this release?
#600

Posting this from windows arm64 device 😥

[martindevans]

@jihadkhawaja It should be possible, it looks like llama.cpp has support for it.

The first step would be to update the build action (see here) to add a windows ARM build. If you're interested in contributing that please feel free to open up an issue and I can give you some tips.

[jihadkhawaja]

I would love to, Will open a PR for that, thanks a lot @martindevans

[LSXAxeller]

I think that you would like to update to the latest llama.cpp commit before releasing, they just got the GLM-4.5 models series PR merged yesterday, and the model seems promising.

EDIT: OpenAI GPT-OSS just released a few hours ago and support merged into llama.cpp.

[martindevans]

That'll have to wait for the next update, this one is basically ready to go. I'm pretty busy right now, but I'll try to find some time this weekend to do another update, hopefully not much has changed in the last 2 weeks!

[Lyrcaxis]

They did a huge PR for gpt-oss 😅 ggml-org/llama.cpp#15091
From a quick scan though not much changed on the mid level except the addition of some enums.