Tencent-backed video company Kuaishou Technology upgraded its Kling AI video generation tool to version 2.1. Users can now upload an image and specify whether they want it animated or recreated.
The company also stated that the upgrade enhances performance, video quality, and production speed.
So far, Kling has over 22 million global users and has generated over 168 million video clips and 344 million images.
The Kling 2.1 version offers 1080p videos
According to Kuaishou, its new Kling 2.1 can produce high-quality 1080p videos in under sixty seconds. Each 1080p video should cost around 100 inspiration credits in Kling’s in-house currency. Additionally, users can get the less refined 1080p option for 35 credits or 720p for 20 credits.
Typically, a user can earn 66 credits per dollar on the platform. The new 2.1 version includes enhanced motion performance and more lifelike physical simulation across all modes.
Kling expects to draw in more users with the upgrade; however, the company had earned a decent amount of revenue even before this. Kling’s earnings report drove its Hong Kong-traded shares up 6% just a day before the version rollout.
Kuaishou released their Kling 2.0 just this April, which Vice President Gai Kun had praised as the most powerful video generation model globally at the time. The model also focused on improving image, movement quality, and the realistic feel of generated clips like the 2.1 version.
The new 2.1 version appears faster and potentially more efficient.
ByteDance upgraded its Doubao chatbot recently
Kuaishou introduced its first version of Kling AI in 2024, following in the footsteps of Open AI’s Sora video model previewed earlier that year
Other tech giants have since rolled out comparable AI services. Among them are ByteDance, Alibaba, and Tencent Holdings. Some AI start-ups like Zhipu AI and Shengshu Tech have joined the effort to develop new AI businesses.
This May, ByteDance upgraded its Doubao chatbot, including a video call function that allows it to interact with users more. Users can now have a video conversation with the AI tech and easily activate the function by turning on their phone’s camera during a voice call. The chatbot can also be used in real-time, for instance, as a tutor with gardening knowledge or an analyst when studying charts.
Earlier, Doubao also enhanced its ability to change images into pixel art. In February, ByteDance also launched its OmniHuman-1 multimodal AI model, which allows users to convert images and sound bites into realistic videos.
In 2024, Zhipu AI launched its AI model, Ying, which generates videos from any text. It can make high-resolution images in a mere 30 seconds. Users can choose how they would like their video presented.
Additionally, Ying offers an image-to-video conversion option, providing new ways to create memes, advertisements, story plots, and short videos.
Moreover, Shengshu AI also introduced its Vidu tool with a “reference to video” feature, which CEO Tang Jiayu argued cuts production time and saves costs. According to the company, Vidu generates a four-second clip in 30 seconds, making it one of the fastest on the market.
At the tool’s inception, Zhang Xudong, product director at Shengshu AI, explained that the company aims to enable users to upload multiple characters and describe scenes, allowing those characters to act within them, much like a film production.
