KW — GGUF Q4_K_M
GGUF Q4_K_M. A 4-bit quantised model file format used by the local inference engine, balancing memory footprint and quality for consumer hardware.
Source: PUMA project documentation + raw benchmark results · Traceability: ANXAG-007,ANXAG-008,ANXAG-009 · Confidence: verified-at-primary-source