Skip to content
AI Atlas
Papercs.CV

BVB: Benchmarking Agentic Video Understanding via Programmatic Reconstruction in Blender

Published 15 Sept 2026arXiv:2609.15478

data quality74

Updated 12 h ago · first seen 14 Sept 2026

paper_01M2HN8YAPN1DVDZ60EC1DA96T

Abstract

Multimodal agents can create complex videos in software such as Blender by coding without relying on diffusion models. Yet video understanding benchmarks still evaluate models mainly through question answering. If an agent truly understands a video, it can reconstruct it programmatically. We introduce BVB, Blender-VideoBench, a benchmark that tests this ability by asking agents to reconstruct real-world videos as animated Blender scenes. To ensure fair comparison, each agent programs the reconstruction through a lightweight harness, Mini-BVB, in an identical sandbox under a shared cost limit. The benchmark renders each reconstruction from its animated camera and evaluates it on two axes: (1) Dual VQA measures how many spatiotemporal facts the reconstruction preserves. (2) Latent Similarity measures how closely the reconstruction matches the source video perceptually. Our overall score, a square-root mean, favors balanced performance. We evaluate 51 configurations from 10 model families and analyze semantic retention, perceptual similarity, reasoning effort, and cost. The best model reaches 88.6 Latent Similarity but retains only 53.7% of the source-correct spatiotemporal answers. Additional reasoning improves visual similarity but does not close this gap in factual accuracy. In a blind study with 15 raters and five configurations, Latent Similarity correlates strongly with human preference. These results show that programmatic reconstruction is a viable test of agentic video understanding, and that semantic retention remains the main challenge.

Authors

Authors 17

Ali VosoughiChao HuangChenliang XuDaiki ShimadaJiayue MengJing BiLuchuan SongMelika FilvantorkamanMingqian FengPinxin LiuQianxiang ShenSusan LiangYicheng WangYolo Y. TangYunzhong XiaoZeliang ZhangZhangyun Tan

Linked names open researcher pages (created from the paper's author list; name-only, no affiliation unless a source states it). Unlinked names have no researcher record yet.

Organizations

Organizations 0

No organization stated. arXiv metadata does not carry affiliations; an organization is linked only when a model card or lab page cites the paper.

Models

Models introduced or described 0

Inbound described_by relations from model cards and documentation.

No model links this paper yet

Model pages link papers through their model cards and documentation; the relation is written only when a source states it.

Datasets

Datasets used 0

No dataset relation recorded.

Benchmarks

Benchmarks used 0

No benchmark relation recorded.

Code

Repositories & frameworks 0

No repository linked.

Timeline

Timeline 2

Full timeline →

Sources

Sources 2

Source documents
SourceDocumentTypeTierLast observedSnapshots
arXiv (Atom API + RSS)rss.arxiv.org/rss/cs.CV feedT1· Official7 h ago4
Hugging Face Hub (public pages, model cards, papers)huggingface.co/papers listingT2· Quality secondary4 h ago30

Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.