AI
Kuaishou Expands Kling 4.0 AI Capabilities for All Users
Kuaishou has released its Kling 4.0 AI model, enabling 30-second video generation, 4K HDR output, and enhanced creative control for all users.
October 16, 2026, 03:22 AM·IT之家
#video-generation
AI
Evaluating AI Video Generation Alternatives After Sora
Following the shutdown of Sora in 2026, a comparative analysis of three alternative AI video generation models evaluates their performance and cost-efficiency.
Хабр
AI
Shengshu Technology Unveils Vidu Q4 Preview With 4K Support
Shengshu Technology has released the Vidu Q4 Preview model, featuring 4K video generation, 16-second clip support, and advanced reference integration for creators.
IT之家
AI
Comparative Performance Analysis of 12 Video AI Models
A comparative study of 12 video generation models evaluated performance, speed, and cost using a standardized prompt, with Luma Ray 3.14 emerging as the most efficient leader.
Хабр
AI
Sber Introduces Kandinsky 6.0 Video for AI-Powered Content Creation
Sber has released Kandinsky 6.0 Video, an AI model for GigaChat that generates Full HD video with synchronized audio.
3DNews - Daily Digital Digest
AI
Sberbank Unveils Kandinsky 6.0 Video for Synchronized Audio-Visual Generation
Sberbank has launched Kandinsky 6.0 Video, an open-source multimodal model that generates videos with synchronized audio in Full HD quality.
CNews.ru
AI
Venice.ai Adds Budget-Friendly Grok Imagine 1.5 Lite Model
Venice.ai has added xAI’s Grok Imagine 1.5 Lite video model, offering generation costs starting at $0.04 per clip.
Crypto Briefing
AI
NVIDIA Unveils Physis-Lang for Physics-Aware Video Generation
NVIDIA researchers launched Physis-Lang, a framework that enhances video models with physics reasoning, outperforming previous benchmarks on the Physics-IQ leaderboard.
MarkTechPost
AI
Google Research Unveils Co-Director for Long-Form AI Video Generation
Google Research introduced four agentic frameworks to improve the coherence and duration of AI-generated videos by addressing consistency and prompt refinement.
MarkTechPost
AI
Google Vids Expands AI Video Tools to Personal Accounts
Google has opened its Vids AI video generation tool to all personal account holders, offering 1080p HD output powered by the Gemini Omni 1.1 Flash model.
Crypto Briefing
AI
Google Unlocks Free 1080p AI Video Generation for All Users
Google has made its AI-powered video generation tool in Google Vids free for all users, allowing 1080p content creation via the Gemini Omni 1.1 Flash model.
Decrypt
AI
TBC Leverages Biological Data for AI Video Optimization
The Biological Computing Co. is partnering with AWS to release a text-to-video model that uses software adapters derived from lab-grown neurons to improve speed and cost efficiency.
The Decoder
AI
Seedance 2.5 Adds Draft Mode to Optimize Video Generation Costs
Volcano Engine has introduced a Draft mode for Seedance 2.5, enabling users to validate creative concepts with 480P previews before generating high-quality 1080P final assets.
IT之家
AI
Generating Long-Form Video on AMD Hardware via ComfyUI
A user successfully implemented a method for generating long-form video on an AMD RX6700XT by using a quantized Wan2.2-I2V-A14B model and a custom frame-stitching workflow in ComfyUI.
Хабр
AI
Fal Unveils H3 Max for High-Speed AI Video Generation
Fal has launched H3 Max, an optimized version of the MiniMax H3 video model that delivers 35 times higher throughput and generates 768p video clips in under three seconds.
Crypto Briefing
AI
Nunchux AI Debuts VC-Attention Kernel to Accelerate Video Diffusion
Nunchux AI has launched VC-Attention, a training-free kernel that accelerates video diffusion transformers by optimizing attention mechanisms on high-end GPUs.
MarkTechPost
AI
Sberbank Introduces Time Adapter for AI Video Dynamics
Sberbank's Kandinsky team has released Time Adapter, an open-source module that enables precise control over temporal dynamics and frame rates in AI-generated videos.
CNews.ru
That is everything