This new one from Qwen should fit though - it looks like that only needs ~30GB of RAM: https://huggingface.co/lmstudio-community/Qwen3-30B-A3B-Inst...
~30GB in Q8 sure, but it's a minimal gain for double the VRAM usage.