Description

SAGE is a specialized EXL3 quantization method designed to optimize large models like DeepSeek-V4.1 for minimal VRAM usage while maintaining high intelligence. It achieves sub-2bit (1.55 BPW) performance that rivals or exceeds 8-bit GGUF benchmarks on hardware like the DGX Spark, maintaining accuracy across benchmarks like HMMT and IFEval.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.