Newest papers

Code2Video-Compass: A Comprehensive Benchmark for General-Purpose Code-Based Video Generation across T2V, I2V, and V2V

Zhenghan Chen, Zekai Shao, Xin Lin, Leting Cheng, Siming Chen · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28416 · cited by 0 works (OpenAlex) · open-access PDF

German

MLLMs (Multimodal Large Language Models) werden benötigt, um vollständige visuelle und zeitliche Leistungsanforderungen in ausführbare Programme zu übersetzen, die vollständige Videos produzieren. Jedoch richten sich die bestehenden Code-basierten Benchmarks hauptsächlich auf spezialisierte Aufgaben und lassen die komplette Videoerstellung in verschiedenen kreativen Kategorien und Eingabemodalitäten unzureichend ausgewertet. Wir führen Code2Video-Compass ein, einen visuellen Holopass, einen allgemeinen Benchmark für die Code-basierte Videoerstellung in vollständige Videos (T2V), Bild-zu-Video (I2V), und Video-zu-Video (V2V) -Einstellungen. Code2Video-Compass umfasst 2,104 Aufgaben, die 10 Kategorien, 24

Machine translation (Meta M2M100); the English original follows.

Abstract

Code-based video generation offers fine-grained control, targeted editing, and component reuse compared with direct video generation. This approach requires multimodal large language models (MLLMs) to translate visual and temporal requirements into executable programs that produce complete videos. However, existing code-based benchmarks primarily target specialized tasks, leaving complete video creation across diverse creative categories and input modalities insufficiently evaluated. We introduce Code2Video-Compass, a general-purpose benchmark for code-based video generation across text-to-video (T2V), image-to-video (I2V), and video-to-video (V2V) settings. Code2Video-Compass comprises 2,104 tasks spanning 10 categories, 24 subtypes, and three complexity dimensions covering construction, dynamic animation, and state dependency. It combines tasks curated from textual instructions and visual references with corresponding T2V, I2V, and V2V tasks derived from complete professional and user-created videos. We propose a dual-level evaluation framework combining 12 objective metrics and five subjective metrics. Objective evaluation covers operational performance, video quality, text-video alignment, and reference fidelity, while subjective evaluation uses an MLLM judge to evaluate instruction following and holistic video quality. Extensive evaluation of 15 recent proprietary and open-weight MLLMs from nine families reveals that even leading models face challenges in reliably generating high-quality complete videos aligned with textual and visual requirements, with performance generally declining as construction, dynamic animation, and state-dependency complexity increases. We anticipate that Code2Video-Compass will advance reliable, coherent, and controllable code-based video generation and inspire broader progress in programmatic visual creation.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Benchmark (surveying), Executable, Modalities, Component (thermodynamics), Modality (human–computer interaction)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange