AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

MMReD: a Cross-Modal Benchmark for Dense Context Reasoning

Maxim Kurkin, Boris Shirokikh, Irina Abdullaeva, Viktoriia A. Chekalina, Andrey Kuznetsov

ICLR 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Unified any-modality language models

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces MMReD, a cross-modal benchmark that tests multi-modal reasoning over dense, information-rich long contexts where simple retrieval is insufficient.

Problem

LLMs and LVLMs with extended context windows still have limited ability to perform complex multi-modal reasoning over long, information-dense contexts.

Approach

A benchmark designed so that answering requires reasoning over dense context rather than simple retrieval, spanning multiple modalities.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add 5 solved in-context examples (from N<=16) (ablation)LLMs 4B-32B · MMReD full datasetrelative improvement: +13.8%Improved
SFT on N<=16 sequences (ablation)7B LLMs · MMReD full datasetrelative improvement: +57.7%Improved
GRPO on N<=16 sequences (ablation)DeepSeek-R1-Distill-Qwen-7B · MMReD full datasetrelative improvement: +27.8%Improved
use GRPO instead of SFT (ablation)7B models · MMReD full datasetrelative improvement: 57.7 -> 27.8Hurt
fine-tune on short sequences (SFT or GRPO) and test on longer contexts (ablation)Qwen2.5-7B, Falcon3-Mamba-7B, R1-Distill-Qwen-7B · MMReD long contextsaccuracyNo clear effect
use text input instead of images (ablation)Qwen2.5-3B-Instruct · MMReD N=128relative gap (%): 6.7 -> -37.5Mixed
use text input instead of images (ablation)GPT-4o-mini · MMReD N=64relative gap (%): 32.5Improved
add 5% perception noise to environment (ablation)LLMs · MMReD NIAH and DCaccuracyHurt
replace visual rooms and characters with symbolic locations and entities (ablation)LLMs · MMReDPearson correlationNo clear effect
increase model parameter count (ablation)Qwen2.5-VL, InternVL2.5, R1-distilled · MMReDaccuracyImproved
use reasoning-specialized LLM (R1-distilled 70B/32B) instead of GPT-4o (ablation)LLMs · MMReDaccuracy: +2-4% initiallyImproved
use code-specialized Qwen2.5-Coder instead of Qwen2.5 (ablation)LLMs · MMReDaccuracyHurt

Experimental setup

Benchmarks: MMReD

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