AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

MMR-Bench: A Comprehensive Benchmark for Multimodal LLM Routing

Hao-Xuan Ma, Guannan Lai, Han-Jia Ye

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Benchmarks and analysis of MLLMs · 8 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
adaptive fusion instead of equal-weight fusion (ablation)KMeansRouter · MMR-Bench full datasetnAUC: +0.3403 nAUC; QNC +inf -> 1.0585Improved
adaptive fusion instead of equal-weight fusion (ablation)MLPRouter · MMR-Bench full datasetnAUC: +0.0033 nAUC; QNC +inf -> 0.9947Improved
adaptive fusion instead of equal-weight fusion (ablation)LinearRouter · MMR-Bench full datasetnAUC: +0.0124 nAUC; QNC 0.9055 -> 0.9701Mixed
adaptive fusion instead of equal-weight fusion (ablation)KNNRouter · MMR-Bench full datasetnAUC: -0.0074 nAUC; Ps -0.0014Hurt
use multimodal features instead of text-only or image-only for routing (ablation)k-means router · MMR-Bench mixed workloadsaccuracy at fixed costImproved
route over MLLM zoo instead of single strongest modelmultimodal router · MMR-Benchcost at matched accuracy: 33% of strongest single model costImproved
use cross-dataset router instead of best single model (OCR)routing policy · OCR scenariopeak score Ps: 0.7062 -> 0.7234Improved
use cross-dataset router instead of best single model (VQA)routing policy · VQA scenariopeak score Ps: 0.7936 -> 0.8012Improved
use cross-dataset router instead of best single model (Math)routing policy · Math scenariopeak score Ps: 0.7592 -> 0.7914Improved
transfer multimodal-trained router to GSM8K text-onlyrouting policy · GSM8Kpeak score Ps: 94.5 -> 96.7Improved
transfer multimodal-trained router to MMLU text-onlyrouting policy · MMLUpeak score Ps: 91.2 -> 92.4Improved
transfer multimodal-trained router to ARC text-onlyrouting policy · ARCpeak score Ps: 65.7 -> 66.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy