Exploring the Trade-Offs: Quantization Methods, Task Difficulty, and Model Size in Large Language Models From Edge to Giant

TL;DR

This study comprehensively evaluates quantization methods on instruction-tuned large language models from 1B to 405B parameters, revealing insights into their robustness, limitations, and the impact of quantization on various tasks.

Contribution

It provides a detailed analysis of four quantization methods across diverse datasets and model sizes, highlighting their effects on performance and task-specific robustness.

Findings

FP8 is the most robust quantization method across tasks.

Smaller models suffer severe accuracy drops at 4-bit quantization.

Quantization amplifies inherent model weaknesses rather than just task difficulty.

Abstract

Quantization has gained attention as a promising solution for the cost-effective deployment of large and small language models. However, most prior work has been limited to perplexity or basic knowledge tasks and lacks a comprehensive evaluation of recent models like Llama-3.3. In this paper, we conduct a comprehensive evaluation of instruction-tuned models spanning 1B to 405B parameters, applying four quantization methods across 13 datasets. Our findings reveal that (1) quantized models generally surpass smaller FP16 baselines, yet they often struggle with instruction-following and hallucination detection; (2) FP8 consistently emerges as the most robust option across tasks, and AWQ tends to outperform GPTQ in weight-only quantization; (3) smaller models can suffer severe accuracy drops at 4-bit quantization, while 70B-scale models maintain stable performance; (4) notably, \textit{hard} tasks do not always experience the largest accuracy losses, indicating that quantization magnifies a model's inherent weaknesses rather than simply correlating with task difficulty; and (5) an LLM-based judge (MT-Bench) highlights significant performance declines in Coding and STEM tasks, though it occasionally reports improvements in reasoning.