# Round-trip Eval: Pre/Post Quant Table — TR-MMLU + MT-Bench + Niche Benchmark

> Source: https://sukruyusufkaya.com/en/learn/fine-tuning-cookbook/ftc-roundtrip-eval-quantization-table
> Updated: 2026-05-14T14:42:57.726Z
> Category: Fine-Tuning Cookbook (Model-by-Model)
> Module: Part X — Quantization Engineering
**TLDR:** Part X capstone: Quantize the same model in bf16, AWQ int4, GPTQ int4, EXL2 4.5bpw, GGUF Q4_K_M, FP8 and compare. TR-MMLU, MT-Bench-TR, niche custom benchmark (Turkish call center sample). Decision matrix: which quant for your use-case?