ByteDance just open-sourced Helios, a 14B video generation model that hits 19.5 FPS on a single Nvidia H100 GPU. It can churn out minute-long videos supporting text, image, and video inputs without quality drifting. Why it matters: bringing near-real-time speeds to minute-scale generation gets open-source video much closer to interactive production.
Google just added Cinematic Video Overviews to NotebookLM, replacing traditional narrated slides with fluid animations and rich visuals. Powered by Gemini 3, Nano Banana Pro, and Veo 3, Gemini acts as a creative director to steer narrative structure and maintain visual consistency. Why it matters: NotebookLM is transforming raw research notes into polished, documentary-style video summaries with zero editing. It is available today in English for Google AI Ultra subscribers (18+) on web and mobile.
LTX Studio just rolled out dubbing and captions features to help enterprise teams localize video content faster. The move expands the AI video platform into multilingual corporate workflows, though details on supported languages and pricing have not been disclosed yet.
Researchers just introduced OmniLottie, a framework that leverages vision-language models to generate lightweight Lottie JSON vector animations directly from text, image, or video prompts. Alongside the model, the team introduced MMLottie-2M, a benchmark dataset of two million annotated vector animations, making automated UI motion design far more practical.
Showlab just open-sourced Kiwi-Edit, a video editing framework that pairs a multimodal LLM with a 5B diffusion transformer to edit clips using plain text or reference images. It handles 720p style transfers, background swaps, and adding or removing objects while keeping temporal consistency. With code, weights, and datasets all released, open-source video manipulation just took a serious step forward.
ByteDance and Stanford researchers just open-sourced Spectrum, a training-free framework that accelerates image and video diffusion models by up to 4.7x. By forecasting future step trajectories with Chebyshev polynomials, it lets models like FLUX.1 and Hunyuan Video generate outputs much faster with virtually no drop in quality.
Researchers just open-sourced WildActor, a video generation framework that keeps an identity's face, body, and clothing consistent across unconstrained motions using only three reference images. Built on Wan2.2-5B, it supports both real people and cartoon characters while releasing the Actor-18M data construction pipeline.
ByteDance and Tsinghua just introduced DreamID-Omni, an AI model that turns reference portraits, text, and voice clips into talking videos. It eliminates lip-sync confusion in multi-person scenes and supports full video face-and-voice editing, with open-source code expected in March.
Google just redesigned its creative platform Flow, folding experimental tools ImageFX and Whisk directly into a single workspace. Creators can now generate assets with the built-in Nano Banana model and immediately use them as ingredients for Veo video generations, alongside new lasso-based natural language precision editing.
Alex Finn fed Cursor's new 'AI records its own demo video' announcement into his OpenClaw agent Henry, which fully rebuilt the feature and recorded a demo in just five minutes. The setup is now embedded in their workflow, automatically attaching demo clips to every pull request.
Sony researchers just introduced MMHNet, a video-to-audio framework that combines hierarchical networks with Mamba architecture to generate realistic audio for clips over five minutes long. It solves the long-video sync puzzle by training on short snippets while generalizing to long scenes, keeping sound effects tightly locked to visual hits and cuts.
Google just updated the Gemini app with a video templates gallery, eliminating the blank-page problem of complex prompting. Users can pick from cinematic styles like Cyberpunk or Action Hero, upload a photo of themselves or a pet to star as the main character, and export in 9:16 vertical video ready for Shorts and Reels.
Researchers just open-sourced VBVR, a massive dataset and evaluation suite aimed at teaching video models how to reason through time and physics. Packed with over a million video clips across 200 curated tasks, it shifts focus away from raw pixel fidelity toward spatial-temporal logic. Why it matters: This marks one of the first major scaling studies in video reasoning, revealing early signs of generalization to completely unseen tasks.
Researchers just introduced SARAH, a real-time conversational agent for VR that actually moves like a human. It tracks your 3D head position to turn toward you, adapt its posture, and hold eye contact at over 300 FPS. Users can even dial in gaze intensity, making VR avatars feel far less like rigid digital puppets.
Researchers just introduced a human-centric video world model that turns tracked head and hand poses from a VR headset into real-time, egocentric interactive video. Running at around 11 FPS, it lets users physically interact with generated environments instead of relying solely on text or keyboard controls. While fidelity is still early, it marks a significant step toward real-time generated XR.
Replit just rolled out Replit Animation, a new tool powered by Gemini 3.1 Pro designed to spin up viral videos in minutes. The company claims users can literally "vibecode" complete animations, and the announcement clip itself was built entirely inside the tool. It signals a aggressive expansion beyond code into AI multimedia creation.
Researchers just introduced VidEoMT, an encoder-only video segmentation model built on standard Vision Transformers. By replacing dedicated tracking modules with a lightweight query propagation mechanism, it matches competitive accuracy while running 5x to 10x faster at up to 160 FPS.
AMC Theatres just pulled an award-winning AI short film called Thanksgiving Day from its pre-show screenings after intense pushback on social media. The short won a festival prize that promised nationwide cinema exposure, but moviegoers revolted over theaters embracing AI. Why it matters: Traditional theatrical spaces remain a minefield for generative AI, with audience sentiment pushing back hard against cinema adoption.
OpenAI just hired Instagram’s longtime partnerships VP Charles Porch to serve as its first-ever VP of global creative partnerships. Porch spent 15 years at Meta wooing icons like Beyoncé and Pope Francis. He is tasked with bridging the trust gap with Hollywood and securing likeness licensing deals for Sora.
Tavus just unveiled Phoenix-4, a real-time human rendering engine that generates full-head motion with millisecond latency and explicit emotion control. Instead of relying on looped video clips and frozen expressions, it actively listens, nods, and shifts moods along with the conversation. Conversational digital humans are finally ditching the uncanny valley.
