FastVideo just released Fast H3 Omni, an open-source acceleration method for MiniMax H3 that cuts sampling steps from 50 to just 8. Built specifically for reference-to-video workflows to keep characters consistent, it slashes latency by up to 12x on top-tier GPUs. Ready-to-use checkpoints for ComfyUI, FP8, and Apple Silicon are already available on Hugging Face.
Gemini Notebook just introduced a co-creation workflow for Slide Decks and Video Overviews, giving users direct control over content and visuals. Instead of jumping straight from prompt to final render, a new Review toggle lets you inspect outlines and producer notes mid-process to tweak scenes. The update rolls out globally to all users over the upcoming week.
Researchers just open-sourced GaussiAnimate, an animation tool that tackles tricky soft-body deformations in garments and animals. Instead of relying solely on a rigid skeleton, it pairs an inner bone structure for motion with flexible outer surface bones to handle complex bends and folds. Why it matters: It removes one of the biggest bottlenecks in 3D character rigging.
Viggle just introduced JST-2, a physics-native world model that turns text prompts or images into interactive 3D scenes in about one second. It tracks persistent physical tokens across space, allowing generated characters to deform and move naturally without manual rigging. It renders directly in the browser via 3D Gaussian splatting, putting full 3D environment generation within reach for everyone.
Cinetic just rolled out an update to generate polished launch videos and UI motion graphics from a single brief. Instead of prompt-to-video hallucination, it uses code to precisely control timelines, sound, and brand assets, making it a sharp companion for coding agents that need quick promo clips.
Mark Wahlberg just voiced his backing for AI in Hollywood at TechCrunch Disrupt, planning to use the tech to enhance stadium scenes in his upcoming Pete Rose biopic. While he opposes AI-generated scripts replacing real actors, he welcomes an AI replica of himself that his children could manage to keep acting indefinitely.
Leonxlnx just open-sourced Earshot, a tool that lets coding agents like Claude Code actually hear what they create. Agents can now inspect audio sync, judge acoustic fit against descriptions, and rewrite synthesis code locally on CPU. It turns tedious trial-and-error sound design into an automated feedback loop.
Sylvester Stallone just revealed he is exploring a Rambo prequel using AI de-aging tech to turn himself back into an 18-year-old high school kid. Following tests inspired by Harrison Ford's digital makeover in Indiana Jones, Stallone confirmed the tech successfully took him decades back in time. Why it matters: Despite fierce Hollywood pushback against AI, aging legends are increasingly using algorithms to keep their most lucrative franchises alive indefinitely.
ElevenLabs just rolled out Instant Soundtracks, a new feature that automatically analyzes uploaded footage to generate original music and frame-accurate sound effects. It adapts moods across cuts, lets users reference existing audio tracks, and separates layers for post-editing via web or API.
Researchers just introduced Möbius, a unified motion-language model that generates, captions, completes, and edits 3D human motion within a single framework. By operating directly on continuous frame-level motion without discrete tokenization or pre-trained motion VAEs, it scales up to 1.7B parameters. Why it matters: It bridges the gap between text prompts and production-ready character animation, allowing creators to describe, generate, and fine-tune complex actions in one place.
Researchers just dropped Any Time, Any Space (ATAS), an open-source framework solving scene drift in long-form AI video. By reconstructing scenes in 3D and storing dynamic spatial memories, cameras can leave a room and return without anything changing. It plugs into models like Wan and MiniMax, with code ready to run locally.
Higgsfield AI just tapped GPT-6 Astra to shatter the typical one-minute ceiling, letting creators generate consistent videos over 10 minutes long. Creative agents can now call external 3D modeling tools to solve continuity headaches, and real-time voice-to-movie rendering is already in the works. Generative video is finally moving from short clips to full-blown production workflows.
Researchers from Tencent and HKU just open-sourced Multix, a world model capable of simulating multiple players interacting in the same environment. It generates individual camera views while syncing player states, cutting computation by 86% with hierarchical communication. Trained on just 4 players, it scales up to 10 without extra training.
OpenAI is in talks to license its Sora video model to a stealth startup backed by Hollywood mogul Jeffrey Katzenberg and former Sora lead Bill Peebles. The startup plans to turn the model into dedicated filmmaking software, while OpenAI takes an equity stake to salvage returns on its heavy compute investments.
Nvidia just open-sourced Stream Render H3, a real-time diffusion system built on MiniMax H3 that replaces traditional game rendering. The engine only tracks physics and outputs flat semantic maps, while the AI generates textures, lighting, and customizable visual styles. Catch is, running it live takes 8 enterprise GP200 GPUs and over 200GB of storage.
Researchers just open-sourced Pure, a 3D reconstruction model that transforms ordinary video footage into interactive dynamic 3D environments. Instead of relying on expensive 3D annotations, it post-trains on regular video by extracting monocular depth and motion cues. Clocking in at just 13.2GB with training scripts included, it makes dynamic 3D mapping practical on local machines.
Researchers just open-sourced MotionSpaceFlow (MSFlow), a lightweight model that turns text prompts into realistic human motions. Weighing around 1.1GB, it runs on consumer GPUs and lets creators pin any joint to a specific coordinate at any frame without task-specific training.
Researchers from Fudan University, Tencent Hunyuan, and Zhejiang University just open-sourced Prism, a framework designed for native 2K joint video-audio generation. Instead of wasting compute across every pixel, Prism uses dynamic sparse attention to focus on sound-producing and rapidly changing visual regions, speeding up training by 2.5x while keeping audio and video tightly synchronized. Preview weights and inference scripts are already available on Hugging Face.
ByteDance introduced DMAD, an adversarial distillation method for rapid visual generation. By training a student model to match real data distributions using a shared-backbone discriminator, it compresses 20-to-30-step video diffusion models like MiniMax-H3 down to four denoising steps while preserving fidelity.
Runway just debuted Praxis-1, an open-weight world action model that adapts to new robots with light fine-tuning, alongside an end-to-end campaign agent called Runway Ads. Co-CEO Cristóbal Valenzuela warned that the US is still obsessed with LLMs while China pours token usage into video models, predicting live physical AI deployments are 12 to 18 months away.
