Cannot load Qwen3.5 GGUF model (unknown architecture)

[ParalelSt]

Description

Tried loading a Qwen3.5 GGUF model and it fails with a generic error:

Failed to load model '/Qwen3.5-9B.Q4_K_M.gguf'

After checking the logs, the actual issue seems to be:

unknown model architecture: 'qwen35'

So it looks like the backend doesn’t support this model yet, but the CLI doesn’t surface that.

Fedora Linux
i5-11400F
RX 6600 (8GB)
16GB RAM

File is definitely there (~5.3GB)
Tried CPU and GPU (--gpu-layers 0 & --gpu-layers more than 0)
Same result every time

Steps to Reproduce

1. Download a Qwen3.5 GGUF model

2. Run:

model load --gpu-layers 0

Platform with bug

Other

Affected platforms

Other

Did you find any workaround?

No response

Relevant log output

llama_model_loader: direct I/O is enabled, disabling mmap llama_model_load: error loading model: error loading model architecture: unknown model architecture: 'qwen35' llama_model_load_from_file_impl: failed to load model