ShengShu shipped Vidu S2 on 15 September: real-time interactive video at 720p, up from 540p. The paper says it outperforms all baselines, and it does. The margin runs from 1.9% to roughly half, depending on which task you point it at.
Anthropic is restoring Claude Fable 5 and Mythos 5 after the US lifted export controls on June 30. Timeline, the new identity and credit rules, and how to resume.
Luma AI released Ray3.2 on June 9, 2026, a video generation update built around frame-level creative control, with up to 16 keyframes per clip, native HDR, and 16-bit EXR export.
ComfyUI v0.24.1 ships Krea 2 Medium Turbo model support, two Bria video background nodes, a Seedance 2.0 1080p artifact fix, and seed control for Flux Erase.
NVIDIA forms Cosmos Coalition with Runway, Black Forest Labs, and LTX to build open video generation infrastructure.
ComfyUI merged native multi-GPU support on May 26, 2026, giving creators with dual or multi-GPU rigs the ability to split image and video generation workloads across all their hardware for the first time.
Adversarial Flow Distillation (AFD) from NUS, LIGHTSPEED, and UCL enables training autoregressive video AI models from closed-source teachers without accessing their weights.
A parameter-free inference technique from UNSW and Griffith University stabilizes lip sync and identity in AI talking-head video without any model retraining.
OpenAI-backed AI feature Critterz arrived at the Cannes market this week without its planned in-festival premiere, after losing Sora as a core tool mid-production. The miss tests the AI-native film pitch and exposes a vendor-risk lesson for working animators.
Runway launched Aleph 2.0 video model and Edit Studio production tool for Pro subscribers.
Google introduced Gemini Omni at I/O 2026, a multimodal model that accepts text, images, audio, and video in any combination and produces or edits video through conversational prompts. Omni Flash is the first public release.
ByteDance Research released Lance, a 3B Apache 2.0 unified multimodal model that handles image and video generation, editing, and understanding in a single framework. Strong VBench and GenEval scores.
NVIDIA and HuggingFace published a full fine-tuning guide for Cosmos Predict 2.5 today, showing how to adapt the 2B-parameter video world model to any domain using LoRA or DoRA on a single GPU.
NVIDIA's SANA-WM is a 2.6 billion-parameter open-source world model that generates 720p, 60-second video with 6-degree-of-freedom camera control on a single GPU.
Three significant open-source models arrived in ComfyUI on May 14, 2026: VOID for video object deletion, BiRefNet for image segmentation, and Gemma 4 for multimodal reasoning.
Google I/O 2026 keynote lands May 19 at 10am PT. Veo 4, Gemini 4, Project Astra, and the Gemini Omni model: here is what creators should watch.
Google's Gemini Omni video model surfaced in the Gemini app on May 11, with demos showing lifelike text rendering, in-chat video editing, and remix tools before the I/O 2026 keynote.
LTX Studio released Flows on May 7, 2026: a visual, node-based system for connecting prompt, image, video, and upscaling steps into a pipeline you can run in one click and reuse across every project.
By mid-2026 ComfyUI hosts thousands of community workflows. Here are 12 that have crossed the production-quality threshold, with download links and the use case each one actually serves.
Runway in 2026 is the most editor-friendly AI video tool. Here is the working video creator's complete guide -- shot generation, camera control, and Premiere or Resolve integration.
Hera Launch ships April 30, 2026 as the YC-backed motion design startup's first focused mode -- generating complete, editable product launch videos from a single text prompt. Code-based animations: every parameter editable after generation.
Novi AI launched its Long Video Agent on April 30, enabling creators to turn scripts into complete 5-minute narrative videos with character consistency and multi-model AI production.
Freepik officially rebranded as Magnific on April 28, 2026, hitting $230M ARR with no outside investment. Inside the model-agnostic aggregator bet, the no-collar economy, and what 1 million paying subscribers means for creative AI.
Picsart's GenAI CLI bundles 130+ image, video, and audio models with native MCP support, turning creative generation into an agent-callable surface.