AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

VLABench: A Large-Scale Benchmark for Language-Conditioned Robotics Manipulation with Long-Horizon Reasoning Tasks

Shiduo Zhang, Zhe Xu, Peiju Liu, Xiaopeng Yu, Yuan Li, Qinghui Gao, Zhaoye Fei, Zhangyue Yin, Zuxuan Wu, Yu-Gang Jiang, Xipeng Qiu

ICCV 2025 · Vision-language and multimodal › Robot action and navigation · approach: Multimodal prompt-driven manipulation · 128 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
add VLA pretraining (vs from scratch) (ablation)OpenVLA · VLABench primitive, seenProgress Score: 3.02 -> 11.74Improved
add VLA pretraining (vs from scratch) (ablation)RDT-1B · VLABench primitive, seenProgress Score: 6.26 -> 15.37Improved
add VLA pretraining (vs from scratch) (ablation)Octo · VLABench primitive, seenProgress Score: 1.02 -> 1.34Improved
use closed-loop control instead of open-loop (64 steps) (ablation)RDT-1B · VLABench primitive, seenProgress Score: 15.37 -> 17.68Improved
shorten open-loop horizon from 64 to 32 steps (ablation)RDT-1B · VLABench primitive, seenProgress Score: 15.37 -> 15.52No clear effect
scale fine-tuning data from 100 to 2000 samples (ablation)Octo, OpenVLA, RDT-1B · Put Box on PaintingProgress ScoreNo clear effect
evaluate on unseen object categories instead of seenRDT-1B · VLABench primitiveProgress Score: 15.37 -> 9.08Hurt
evaluate on unseen object categories instead of seenOpenVLA · VLABench primitiveProgress Score: 11.74 -> 7.93Hurt
switch to common-sense instructions for unseen objectsRDT-1B · VLABench primitive, unseenProgress Score: 9.08 -> 3.61Hurt
switch to common-sense instructions for unseen objectsOpenVLA · VLABench primitive, unseenProgress Score: 7.93 -> 5.11Hurt
add visual perception (GPT-4V) to Voxposer (ablation)Voxposer · VLABench workflowsProgress ScoreMixed

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