On September 19th, Kuaishou Technology announced a significant global upgrade to its KeLing AI model, introducing version 1.5. The new version has achieved notable improvements in various aspects, including video quality, dynamic effects, aesthetic performance, motion plausibility, and semantic understanding.
The upgraded version 1.5 model supports direct output of 1080p high-definition video in high-quality mode, marking a major leap forward in video clarity and texture. Compared to the previous version 1.0, the version 1.5 model has significantly improved in terms of image quality, dynamic effects, and text responsiveness, with internal evaluations showing an overall improvement of 95%.
In terms of dynamic quality, the new model enhances the dynamic effects of videos through more realistic physical representations and naturally smooth movements. Additionally, the version 1.5 model can respond to more complex text descriptions, demonstrating powerful image-to-video understanding capabilities.
To further enhance the precision control of video generation, KeLing AI introduced a new "Motion Brush" feature. This feature allows users to specify motion trajectories for elements in an image, achieving precise motion control.
In addition to the above upgrades, KeLing AI has introduced a series of other features, including support for generating multiple videos at once, increasing video duration, and supporting image quality enhancement. The official also released a user guide to help users better master the various functions of KeLing AI.
Since its launch in June this year, this marks the 9th iteration upgrade for KeLing AI. In July, KeLing AI announced the official release of the international version 1.0, followed by the launch of a global membership system. To date, KeLing AI has accumulated over 2.6 million users, generating more than 27 million videos and 53 million images, establishing itself as a leader in the global AI video generation field.