Saluki 27B: The 2-Bit AI Model Outperforming Its Predecessor in Tool Calling
Researchers have released Saluki 27B, a highly compressed 2-bit GGUF version of the Qwen3.8-27B model that requires just 7.89 GB of storage. Despite its drastically reduced size compared to the 54 GB original, it demonstrates superior tool-calling capabilities. However, the compression compromises its performance in advanced competition mathematics and complex reasoning tasks.
Source: MarkTechPost