base_model: nvidia/OpenReasoning-Nemotron-7B
EXL3 quantization of OpenReasoning-Nemotron-7B, 6 bits per weight.