AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

OpenVision: A Fully-Open, Cost-Effective Family of Advanced Vision Encoders for Multimodal Learning

Xianhang Li, Yanqing Liu, Haoqin Tu, Cihang Xie

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability analysis · 9 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 patch size 8 instead of 16 (ablation)OpenVision Ti · LLaVA-1.5 TextVQAaccuracy: 50.2 -> 54.6Improved
use patch size 8 instead of 16 (ablation)OpenVision S · LLaVA-1.5 TextVQAaccuracy: 54.3 -> 59.3Improved
use patch size 8 instead of 16 (ablation)OpenVision B · LLaVA-1.5 TextVQAaccuracy: 57.9 -> 61.2Improved
add auxiliary text decoder (ablation)OpenVision · multimodal benchmarksaccuracyImproved
use synthetic captions instead of original DataComp-1B captions (ablation)OpenVision · multimodal benchmarksaccuracyImproved
add 336 stage after 224 (resolution progression) (ablation)OpenVision L/14 · Open-LLaVA-Next TextVQAaccuracy: 64.4 -> 65.7 -> 68.3Improved
double higher-resolution fine-tuning samples (ablation)OpenVision L/14 · Open-LLaVA-Next ChartQAaccuracy: 66.1 -> 68.0Improved
scale Stage 3 data from 665K to 3M (ablation)OpenVision-Smol · Smol-150M ChartQAaccuracy: 10.8 -> 31.3Improved
scale input resolution from 384 to 768 (ablation)OpenVision-Smol · Smol-150M TextVQAaccuracy: 36.3 -> 40.6Improved
scale Stage 2 data from x1 to x8 (ablation)OpenVision-Smol · Smol-150M SEED-Benchaccuracy: 50.0 -> 45.1No clear effect
scale encoder from L/14 to H/14OpenVision · LLaVA-1.5 OCR-Benchscore: 315 -> 330Improved
use OpenVision B/16 instead of OpenAI CLIP B/16 as encoderQwen2.5-0.5B LLaVA-1.5 · TextVQAaccuracy: 33.5 -> 34.8Improved

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