Error: unable to load model Caused by: null result from llama cpp

[wighawag]

Hey, not sure if llama-cpp-rs is supposed to run any gguf model but i tried with the latest ministral model from unsloth : huggingface.co/unsloth/Ministral-3-3B-Instruct-2512-GGUF but it failed

I executed

cargo run --release --bin simple -- --prompt "The way to kill a linux process is" local ~/Downloads/Ministral-3-3B-Instruct-2512-UD-Q4_K_XL.gguf

and got the following error

Error: unable to load model

Caused by:
    null result from llama cpp

any idea?

[MarcusDunn]

If you can successfully do it via llama.cpp, but this library fails, I would consider it a bug.

It's possible an update of the bindings will fix this (we may not be on a llama.cpp version that supports the latest ministral).

PRs welcome! binding updates are generally very simple to do.

[wighawag]

I can run it via

./llama.cpp/llama-cli  -m ~/Downloads/Ministral-3-3B-Instruct-2512-UD-Q4_K_XL.gguf -c 32684 -ngl 99 -fa on --threads 10

so i guess this is a binding update needed.