Can confirm here as well. Running ilintar/qwen3.8-flash-next-gguf-strix-halo (IQ4) on pwilkin/strix-llama.