AI

Non-uniform (mixed-precision) GGUF quantizations of Qwen3.8-27B from the IST Austria DAS Lab, built with GSQ and RCO. The smallest file is 8.4 GB and already beats BF16 on zero-shot tasks; the 11.8 GB variant is task-lossless. Standard GGUF that runs in llama.cpp, Ollama and LM Studio.

You don’t have to pick just one machine to run your model on. If you have two PCs with GPUs on the same network, llama.cpp can use all of them…