OpenCodePapers

open-domain-subject-to-video-on-opens2v-eval

Image to Video GenerationOpen-Domain Subject-to-Video

Results over time

Click legend items to toggle metrics. Hover points for model names.

Leaderboard

Paper	Code	Total Score	Venue	Aesthetics	Motion	FaceSim	GmeScore	NexusScore	NaturalScore	ModelName	ReleaseDate
Identity-Preserving Text-to-Video Generation by Frequency Decomposition	✓ Link	0.5446	Close-Source	0.4460	0.4160	0.4010	0.6620	0.4592	0.7906	Kling 1.6	2024-11-26
VACE: All-in-One Video Creation and Editing	✓ Link	0.5287	Open-Source	0.4721	0.1502	0.5509	0.6727	0.4420	0.7278	Wan2.1-VACE-14B	2025-03-10
Phantom: Subject-consistent video generation via cross-modal alignment	✓ Link	0.5232	Open-Source	0.4639	0.3342	0.5148	0.7065	0.3743	0.6866	Phantom-Wan-14B	2025-02-16
Phantom: Subject-consistent video generation via cross-modal alignment	✓ Link	0.5071	Open-Source	0.4667	0.1429	0.4855	0.6942	0.4244	0.7026	Phantom-Wan-1.3B	2025-02-16
SkyReels-A2: Compose Anything in Video Diffusion Transformers	✓ Link	0.4961	Open-Source	0.3940	0.2560	0.4595	0.6454	0.4377	0.6722	SkyReels-A2-Wan2.1-14B-Preview	2025-04-03
Identity-Preserving Text-to-Video Generation by Frequency Decomposition	✓ Link	0.4888	Close-Source	0.4687	0.2470	0.3080	0.6921	0.4541	0.6979	Pika 2.1	2024-11-26
MAGREF: Masked Guidance for Any-Reference Video Generation	✓ Link	0.4793	Open-Source	0.4502	0.2181	0.3083	0.7047	0.4304	0.6949	MAGREF-480P	2025-05-29
Identity-Preserving Text-to-Video Generation by Frequency Decomposition	✓ Link	0.4759	Close-Source	0.4147	0.1352	0.3511	0.6757	0.4355	0.7144	Vidu 2.0	2024-11-26
VACE: All-in-One Video Creation and Editing	✓ Link	0.4553	Open-Source	0.4824	0.1883	0.2058	0.7126	0.3795	0.7178	Wan2.1-VACE-1.3B	2025-03-10
VACE: All-in-One Video Creation and Editing	✓ Link	0.4395	Open-Source	0.4727	0.1203	0.1658	0.7138	0.4004	0.7056	Wan2.1-VACE-1.3B-Preview	2025-03-10