AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

MIEB: Massive Image Embedding Benchmark

Chenghao Xiao, Isaac Chung, Imene Kerboua, Jamie Stirling, Xin Zhang, Márton Kardos, Roman Solomatin, Noura Al Moubayed, Kenneth Enevoldsen, Niklas Muennighoff

ICCV 2025 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 19 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
use MLLM-based embedding models instead of CLIP-style modelsMIEB · MIEB Full overallMean (all tasks): Voyage 62.5 / E5-V 58.2 vs best CLIP-style siglip 57.1Improved
MLLM-based embedding versus CLIP-style on linear probing and zero-shot classificationE5-V · MIEB Zero-shot Classificationaverage accuracy: E5-V 50.0 vs siglip-so400m 70.8Hurt
scale OpenCLIP model size from ViT-L (430M) to ViT-H (990M) (ablation)OpenCLIP · Document UnderstandingnDCG@5: 36.26 -> 40.41Improved
use higher-quality data (DataComp) with fewer seen samples (ablation)CLIP ViT-L · Document UnderstandingnDCG@5: 36.26 -> 38.64Improved
use text-supervised CLIP models versus vision-only self-supervised models on visual STSVision encoders · Visual STSSpearman: EVA-CLIP-bigE-14-plus 71.99 vs DINO-v2 12.98, MOCO-v3 14.31Improved
use few-shot linear probing with k from 8 to 256 (ablation)MIEB linear probing · fine- and coarse-grained classificationmodel rankingNo clear effect
use a strong multilingual text encoder (frozen BGE-M3 in VISTA, multilingual SigLIP)Image-text embedding models · Multilingual retrievalaverage score across languagesImproved
reduce MIEB to MIEB-lite (51 tasks)MIEB · E5-V runtimeH100 GPU hours: 264.0 -> 46.4Improved
predict MLLM benchmark performance from vision-encoder Visual STS scoreVision encoders · 16 MLLM benchmarkscorrelation: over 99% correlation on OCR tasksImproved

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