Great for home assistant!
From quick testing it seems to integrate quite well with home assistant.
Also its the first model which seems to not only speak german, but also does so in a quite natural way Im impressed.
Huge step up from the original qwen3 next
Runs with 200k context on 3x rtx 3090 AWQ quant with pipeline parallel on vllm
Cheers :)
Moltbook long context testing on the way haha
How do you run this at home?
VLLM
But I did not mean that in the literal sense, but this: https://www.home-assistant.io/
Anyways Im mind blown by this model, it works great and its so fluent!
Quite nice to talk to actually.
Instant favourite, thanks guys :)
It's also good for OpenClaw. I didn't notice any significant difference between it and MiniMax or GLM in common tasks. Considering its parameters, local deployment for OpenClaw may very suitable.
It's also good for OpenClaw. I didn't notice any significant difference between it and MiniMax or GLM in common tasks. Considering its parameters, local deployment for OpenClaw may very suitable.
Im testing this right now.