BWhat changed
Integrated a high-compression video VAE, bilingual prompt encoders, full spatiotemporal attention and staged flow-matching/preference training, accompanied by an openly released video model and evaluation set.
Scientific scope
Integrated a high-compression video VAE, bilingual prompt encoders, full spatiotemporal attention and staged flow-matching/preference training, accompanied by an openly released video model and evaluation set. One original system and companion evaluation resource. DiT, Flow Matching and DPO are inherited; the source explicitly reports failures on complex action and physical consistency. No world-simulation or universal quality claim.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. One original system and companion evaluation resource. DiT, Flow Matching and DPO are inherited; the source explicitly reports failures on complex action and physical consistency. No world-simulation or universal quality claim.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model
- Step-Video Team: compressed video generation and preference training — official research record
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction