xAI just released the Grok Imagine API, rolling out unified models to turn text prompts and static images into dynamic video with native audio. Built specifically to tackle rendering lag and heavy compute costs, it also enables localized object editing and stylistic scene transforms. Why it matters: by optimizing latency and price, xAI aims to let creative teams test parallel ideas without painful iteration bottlenecks.
Researchers just released TeleStyle, an open-source model that restyles both images and videos using reference pictures or text prompts. Built on Qwen-Image-Edit, it pairs curriculum learning with a dedicated video-to-video module to keep content and motion consistent across frames.
Tencent just introduced ScriptAgent, a multi-agent framework that turns dialogue into cinematic videos. Three specialized agents handle scripting, orchestrate video models like Sora2-Pro and Veo3.1 across scenes, and critique visual quality. Why it matters: It universally boosts script faithfulness across top models, tackling one of generative video's biggest pain points—long-horizon narrative coherence.
Google DeepMind just debuted 'Dear Upstairs Neighbors' at the Sundance Film Festival, an animated short co-created with animation veterans. Instead of relying on text prompts, researchers built custom video-to-video workflows and local refinement tools powered by Veo and Imagen. Why it matters: It shows a practical path where professional animators keep precise creative control instead of wrestling with AI randomness.
Luma AI just launched Ray3.14, a major upgrade to its reasoning video model that outputs native 1080p footage. It delivers four times faster generation speeds at one-third of the previous per-second cost, cutting down on flicker and drift so studios can use raw generations directly in production.
Runway just rolled out image-to-video for Gen-4.5, claiming only 10% of participants could distinguish its outputs from real footage. With the model topping Artificial Analysis video leaderboards, Runway warned that hyper-realism has hit a societal tipping point where seeing is no longer believing.
YouTube just announced that creators will soon be able to generate Shorts using their own AI likeness, alongside text-to-game creation tools. To prevent AI slop and identity theft, the platform is pairing the feature with strict likeness-management controls. It's rolling out later this year.
Researchers just introduced VideoMaMa, a model that uses generative video diffusion priors to turn coarse segmentation masks into pixel-accurate alpha mattes. Even with only synthetic training, it handles in-the-wild video effortlessly and powers a massive 50K-video dataset. All models and data are slated for public open-source release.
Researchers just open-sourced Motion 3-to-4, a framework that synthesizes coherent 4D dynamic objects from a single monocular video. By decoupling static 3D shapes from motion latent representations, it recovers per-frame vertex trajectories and can even transfer character motion into entirely new videos.
Researchers just introduced OmniTransfer, a Wan 2.1-based framework that unifies spatial appearance and temporal video transfer. It replicates complex VFX, fluid character motion, and cinematic camera moves directly onto new clips without manual pose or trajectory extraction.
Lightricks just rolled out an audio-to-video feature that lets creators generate video directly from speech, music, or sound effects. The model understands beats and rhythm to keep visuals in sync with the track, making it easy to generate music clips, b-roll, and digital avatars. Basically, audio is becoming the new visual prompt.
Adobe just launched an on-device AI Object Mask tool for Premiere, letting editors track complex moving subjects with a simple hover and click. Redesigned shape masks track up to 20x faster, and editors can directly import assets from Firefly Boards. After Effects 26.0 also launched with native 3D parametric meshes and variable font animation.
Overworld just dropped Waypoint-1, an interactive video diffusion model you can explore in real time using a keyboard and mouse. Powered by its new WorldEngine inference library, the 2.3B model hits up to 60 FPS on an RTX 5090 with near-zero control lag. Visual consistency and hallucinations are still rough around the edges, but you can already test the demo or run it locally.
Remotion's Agent Skills now lets AI coding agents like Claude Code script, animate, and render polished motion graphics entirely from plain text prompts. In a recent demo, Claude generated React components, orchestrated 3D camera flips, and delivered an 8-second promo without touching After Effects. Basically, product demos and UI animations can now be vibe-coded on demand.
Researchers just open-sourced CoDance, a new animation model that can animate multiple subjects from a single, potentially misaligned pose sequence. It breaks rigid spatial constraints and accurately rebinds motion to target characters using text prompts and masks. All code and model weights are now publicly available.
Researchers just introduced FlowAct-R1, a streaming humanoid video generation model capable of synthesizing infinite-duration video in real time. It achieves 25fps at 480p with a 1.5-second first-frame latency while maintaining natural behaviors. Why it matters: This drastically cuts latency for responsive AI avatars and seamless live streaming.
Columbia Engineering researchers just built a flexible silicone robot face powered by 26 tiny motors that learned to lip-sync by watching YouTube videos instead of following rigid rules. After practicing its own facial mechanics in a mirror, its vision-to-action model beat five existing systems across 11 languages. Why it matters: Paired with ChatGPT, it could finally cure the stiff, uncanny lip-syncing that has plagued humanoids for years.
Google just updated its Veo 3.1 video model with native 9:16 portrait outputs and up to 4K resolution. The refreshed Ingredients to Video tool lets you feed in reference photos to maintain character and background consistency across scenes without face distortion, tailored directly for short-form creators.
PixVerse just launched R1, an interactive AI world model that streams continuous 1080P video in real time. Instead of waiting for fixed clips to render, you can inject prompts mid-stream to spawn cars or trigger background explosions on the fly. It is still glitchy and prone to hallucinations, but it shifts AI video from static outputs into live simulations.
Researchers just introduced RigMo, a framework that learns both skeletons and motion directly from raw 3D mesh sequences without human rigging annotations. It handles humans, animals, and even objects like fans with fewer artifacts than previous methods. Why it matters: It turns static or deforming 3D meshes into game- and film-ready animatable assets automatically.
