AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

CoT-VLNBench: A Benchmark for Visual Chain-of-Thought Reasoning in Vision-Language-Navigation Robots

Xiao Zhao, Chang Liu, Ruiteng Ji, Zheyuan Zhang, Mingxu Zhu, Linna Song, Zhe Ren, Luo Qingliang, YuHang Gao, Zhaolong Du, Chufan Guo, Kuifeng Su

AAAI 2026 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability analysis

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 CoT-VLNBench, a benchmark for visual chain-of-thought reasoning in vision-language-navigation robots, motivated by the lack of vision-language task support in robot-centric datasets.

Problem

Existing robot-centric datasets focus on traditional 3D tasks such as perception and prediction and lack adequate support for vision-language navigation (VLN) reasoning in embodied settings.

Approach

A benchmark targeting visual chain-of-thought reasoning for vision-language-navigation robots, aimed at interpretable navigation in complex environments. Full method details are unavailable from the abstract.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add CoT reasoning to vision-language training (ablation)Qwen2.5-VL 7B · CoT-VLNBench valavg L2 error (m): 0.33 -> 0.29Improved
Add CoT reasoning to vision-language training (ablation)All four VLMs on average · CoT-VLNBench valavg L2 error (m): 0.37 -> 0.34Improved
Add visual input to language-only training (ablation)All four VLMs on average · CoT-VLNBench valavg L2 error (m): 0.42 -> 0.37Improved
Add visual input and CoT to language-only model (ablation)Qwen2.5-VL 7B · CoT-VLNBench valL2 error at 3s (m): 0.62 -> 0.47Improved
Add CoT reasoning (ablation)InternVL3-8B · CoT-VLNBench valavg L2 error (m): 0.34 -> 0.31Improved
Add CoT reasoning (ablation)Janus-Pro 7B · CoT-VLNBench valavg L2 error (m): 0.42 -> 0.39Improved
Add CoT reasoning (ablation)LLaVA-1.6 7B · CoT-VLNBench valavg L2 error (m): 0.40 -> 0.35Improved
Use CoT-VLN (Qwen2.5-VL) instead of SparseDriveNon-VLM trajectory predictors · CoT-VLNBench valavg L2 error (m): 0.42 -> 0.29Improved
Use CoT-VLN (Qwen2.5-VL) instead of DiffusionPlannerNon-VLM trajectory predictors · CoT-VLNBench valavg L2 error (m): 0.37 -> 0.29Improved
Use Qwen2.5-VL backbone instead of LLaVA-1.6CoT-VLN · CoT-VLNBench valGPT-4.1 CoT score vs L2: score 66.48 -> 75.67; L2 0.35 -> 0.29Mixed

Experimental setup

Benchmarks: CoT-VLNBench

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