Is there anyone here who is using OpenHands with models served from a DGX Spark? I currently have a MinMax-M2.7 and am looking to run GLM5.2 on the machine, but the response time is slow. I was switching from using qwen3-coder-30b so I can run prompts that can connect with large codebases. Please let me know I will appreciate any help.

these GLM 5.2 and minimax M 2.7 are very huge for the dgx spark , like these are 500B models which is huge , and for usable speeds use models under 120B and since you are using qwen 3 30B coder , you can use Ornith-1.0-35B which is far more better and can reach opus level in some benchmarks , feel free to ping me for further more clarification

Alright thank you for the help