MBQ (LLM quantization): superseded — cited as a baseline and beaten by newer methods. 2 paper(s) critique it, 2 beat it on benchmarks — #18 of 80 most-superseded. Sub-problem: cluster led by GPTAQ. Newer alternatives in the same sub-problem include SplitQ, MARR, Quantization-aware Integrated Gradients (QIG), SPEED-Q, Quant-dLLM.

Superseded baseline#18 of 80 most-superseded

MBQ

MBQ: Modality-Balanced Quantization for Large Vision-Language Models

LLM quantization · first seen Dec 27, 2024

superseded — cited as a baseline and beaten by newer methods

2 papers critique it · 2 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites MBQ as a baseline.

“Existing approaches~li2025mbq typically learn a single transformation shared by both modalities across all channels, where the cross-modal heterogeneity can severely distort the optimization objective.”
— Breaking Modality Heterogeneity in Low-Bit Quantization for Large Vision-Language Models
“these methods either require expensive parameter fine-tuning xie2024qslaw, specialized manipulation at inference time yu2025mquant, or rely on a suboptimal grid search li2025mbq, failing to offer an efficient and effective solution for both calibration and inference”
— VLMQ: Efficient Post-Training Quantization for Large Vision-Language Models via Hessian Augmentation

Beaten on benchmarks

Head-to-head results where a newer method reports beating MBQ. Values are copied from the source paper's tables — verify against the cited paper.

SplitQ beats MBQ · Avg. [W4A8]
70.4 vs 63.5
Breaking Modality Heterogeneity in Low-Bit Quantization for Large Vision-Language Models
SplitQ beats MBQ · Avg. [W4A4]
69.6 vs 4.2
Breaking Modality Heterogeneity in Low-Bit Quantization for Large Vision-Language Models
QIG (Ours) beats MBQ · Avg. [LLaVA-onevision-7B W3A16]
72.04 vs 70.44
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
QIG (Ours) beats MBQ · Avg. [LLaVA-onevision-7B W4A8]
70.23 vs 70.16
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
QIG (Ours) beats MBQ · Avg. [InternVL2-8B W3A16]
72.31 vs 72.20
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
QIG (Ours) beats MBQ · Avg. [InternVL2-8B W4A8]
72.04 vs 71.38
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
QIG (Ours) beats MBQ · Avg. [Qwen2-VL-7B W3A16]
70.30 vs 70.15
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
QIG (Ours) beats MBQ · Avg. [Qwen2-VL-7B W4A8]
67.77 vs 67.48
Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients

Newer alternatives

Recent methods in the same sub-problem, not yet superseded in the knowledge base.