AI Paper Atlas

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

CapNav: Benchmarking Vision Language Models on Capability-conditioned Indoor Navigation

Xia Su, Ruiqi Chen, Benlin Liu, Jingwei Ma, Zonglin Di, Ranjay Krishna, Jon Froehlich

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN · 4 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
enable thinking mode (9 matched pairs) (ablation)VLMs on CapNav · CapNavCapNav score: +6.87%Improved
enable thinking mode (ablation)VLMs on CapNav · CapNavinference time: 14.94 s -> 123.94 sHurt
enable thinking mode (ablation)Intern-VL3.5-8B · CapNavCapNav score: 41.72 -> 53.72Improved
enable thinking mode (ablation)Keye-VL-1.5-8B · CapNavCapNav score: 47.28 -> 50.47Improved
enable thinking mode (ablation)Doubao-Seed-1.6 · CapNavCapNav score: 61.91 -> 62.12No clear effect
enable thinking mode (ablation)Intern-VL3.5-14B · CapNavreasoning validity: 25.18 -> 24.97No clear effect
increase input frames (16, 32, 64, 1 FPS) (ablation)VLMs on CapNav · CapNavCapNav scoreMixed
LoRA fine-tune on augmented 13K tasksQwen3-VL-8B-Instruct · CapNavCapNav score: 45.26% -> 55.18%Improved
LoRA fine-tune on augmented 13K tasksQwen3-VL-8B-Instruct · CapNavreasoning validity: 0.30 -> 0.25Hurt
tighten mobility constraints (humanoid vs adult human)VLMs on CapNav · CapNavmean CapNav score: 57.83% -> 39.12%Hurt
use spatial-reasoning models instead of general VLMsSpatial-MLLM-4B · CapNavCapNav score: 30.15Hurt
vary CapNav score weights (0.5 on one component) (ablation)model ranking · CapNavKendall tau: 0.909No clear effect

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