Kunlun Tech’s SkyReels-V4 has ranked No.1 globally within the “text-to-video (with audio)” class on the Synthetic Evaluation benchmark, surpassing fashions together with Kling 3.0, Google Veo 3.1, and OpenAI Sora 2. The mannequin beforehand ranked second following its preview launch in February.
The newest improve introduces two key technical enhancements. First, a full-modality reinforcement studying framework integrates a semantic reward mannequin with curriculum studying, enabling the system to generate 15-second 1080p movies whereas sustaining logical coherence in scene development.
Second, the mannequin provides keyframe and grid-based visible reference capabilities, permitting customers to add as much as 9 photos. This permits stronger consistency in character look and scene model throughout generated sequences.
Following its rise to the highest of the rankings, SkyReels-V4 has opened API entry, supporting text-to-video, image-to-video, multimodal technology, in addition to video enhancing and restoration.
Supply：QbitAI
Elevate your perspective with NextTech Information, the place innovation meets perception.
Uncover the most recent breakthroughs, get unique updates, and join with a world community of future-focused thinkers.
Unlock tomorrow’s developments at the moment: learn extra, subscribe to our e-newsletter, and turn out to be a part of the NextTech group at NextTech-news.com
