Recently, researchers from ByteDance Research Institute and Tsinghua University jointly released a new study, pointing out that current AI video generation models, such as OpenAI's Sora, while capable of creating stunning visual effects, have significant flaws in understanding basic physical laws. This study has sparked widespread discussion on the capabilities of AI in simulating reality. The research team tested AI video generation models under three different scenarios: predictions under known patterns and predictions under unknown patterns.

TikTok recently announced that its AI-driven video generation tool, Symphony Creative Studio, is now fully available to all TikTok business users. This tool will provide advertisers and content creators with an unprecedented creative experience. This intelligent tool, which integrates video generation, transformation, and expansion functions, aims to help advertisers and creators overcome the bottleneck between creativity and production, making the video creation process simpler and more efficient. Users can access the full suite of features at no additional cost.

On the occasion of Vidu's launch anniversary, Shengshu Technology is proud to announce the release of the new version Vidu 1.5, which achieves world-leading breakthroughs, especially in understanding diverse inputs and overcoming the 'consistency' challenge. The launch of Vidu 1.5 marks a new 'contextual' era for visual models, accelerating the arrival of Artificial General Intelligence (AGI). From its global debut, Vidu has demonstrated the ability to generate character consistency.

Today, the Zhipu Technology team released its latest video generation model, CogVideoX v1.5, and made it open source. This version is another significant advancement in the CogVideoX series launched by the Zhipu Technology team since August. The update greatly enhances video generation capabilities, including support for 5-second and 10-second video lengths, 768P resolution, and the ability to generate 16 frames. Additionally, the I2V (Image-to-Video) model supports arbitrary aspect ratios, further improving the understanding of complex semantics.

Beijing Zhizhu Huazhang Technology Co., Ltd. announced that its CogVideoX series models have launched the latest version - CogVideoX v1.5, which is now open source. Since its release in early August, this series of models has become a leader in the video generation field due to its industry-leading technology and developer-friendly features. The new version, CogVideoX v1.5, has been significantly upgraded, enhancing video generation capabilities and now supports 5/10 second, 768P, 16-frame video generation.

Recently, Kling introduced a brand new "Custom Models" feature, a significant update that will thoroughly address the technical challenges of character consistency in AI video generation. Based on feedback from early testing users on the platform, this feature allows creators to upload 10 to 30 different video clips for model training, with each clip requiring at least 10 seconds. Unlike traditional models that rely solely on image training, KLING innovatively utilizes video materials as the training basis and even supports using AI-generated videos created by Kling for training.

Generating high-quality, temporally continuous videos requires a significant amount of computational resources, especially for longer time spans. The latest Diffusion Transformer model (DiTs) has made significant progress in video generation, but its reliance on larger models and more complex attention mechanisms has resulted in slower inference speeds, exacerbating the challenge. To address this issue, researchers at Meta AI have proposed a training-free method called AdaCache to accelerate video DiTs.

Beijing Aishi Technology Co., Ltd. has announced the global release of its latest version of the video generation large model, PixVerse V3. This version has made significant improvements in its underlying model capabilities, offering multimodal generation abilities and a wealth of AI features, including creative templates, lip synchronization, story continuation, and style transformation.

As a new wave of competition arises in the global AI video generation field, Chinese tech company MiniMax's 'Haibao AI' has made a strong impression with its remarkable performance. Data shows that this AI product, which was launched in April of this year, achieved a staggering 867.41% increase in visit volume in September, rapidly becoming the top product in global AI visit growth rankings. The core competitiveness of this product lies in its underlying technological support. MiniMax built Haibao AI based on the trillion-parameter MoE model abab-6.5, allowing it to compete with GPT-4.

Kuaishou, Peking University, and Beijing University of Posts and Telecommunications have joined forces to unveil a groundbreaking technological achievement—the Pyramid-Flow high-definition video model. This open-source project has made significant breakthroughs in the field of AI-generated video, bringing new possibilities to the industry. The Pyramid-Flow model demonstrates remarkable capabilities, generating high-quality videos of up to 10 seconds in length, with a resolution of 1280x768 and a frame rate of 24fps, using only text input. It excels in light and shadow effects, motion continuity, overall image quality, and text interpretation.

Recently, Kuaishou, along with research teams from Peking University and Beijing University of Posts and Telecommunications, jointly open-sourced an ultra high definition video generation model named Pyramid-Flow. This model can generate videos of up to 10 seconds long, with a resolution of 1280x768 and 24 frames per second, producing high-quality results with excellent lighting effects, action consistency, and overall video quality. Pyramid-Flow operates differently from existing video diffusion models, which typically run at full resolution; while this approach allows for...

Recently, AI video generation technology has garnered significant attention. To promote the development of this emerging field, Runway announced the establishment of a new film fund, planning to invest 5 million dollars to support the production of up to 100 original films. This is no small amount, especially for directors looking to leverage AI in their creative processes, as it presents a rare opportunity. Runway's '100 Film Fund' aims to finance projects that can apply their generative video models in the creative process.

At the Baidu Cloud Intelligence Conference held today, Tang Jiayu, co-founder and CEO of Shengshu Technology, announced that Vidu, the first domestic video large model, officially opens its API and integrates with Baidu Intelligent Cloud's Qianfan Large Model Platform, becoming the first video large model on this platform. Vidu's self-developed video generation technology boasts advantages such as high dynamism, diverse styles, and exceptional reasoning. It has also launched the 'Subject Reference' feature globally, effectively solving the problem of consistent video model generation. Since its launch at the end of July, Vidu has received thousands of enterprise user integrations.