Google's new multimodal generative model announced at Google I/O 2026 (May 19), rolling out free on YouTube Shorts and the YouTube Create app. Creates and edits video from any combination of text, image, audio, and existing video. Key capabilities: multi-turn conversational editing — describe changes in plain language and the model holds character and physics consistent across edits; reference-input generation — upload your own image, video, or drawing and describe the scene you want; native SynthID watermarking on every output (relevant for EU AI Act compliance deadline December 2, 2026); and avatar creation, currently in limited testing rollout. For teams already on Google AI Pro or Ultra, this is a significant capability expansion at no additional cost. Currently outputs at 720p per community testing. Developer API described as 'coming in the coming weeks' — not yet live. Competes directly with Runway, Kling, HeyGen, and Pika — tools that carry per-generation or subscription fees.

Professional AI video generation platform that received a landmark update on May 21, 2026 with the release of Aleph 2.0 and Edit Studio — described by the community as 'Photoshop for video.' Unlike standard AI video generation that creates new clips from scratch, Aleph 2.0 makes localized, targeted edits to existing video: change one element while the rest of the frame stays intact, propagate a single edit across an entire multi-shot cut sequence without touching every clip manually, and use a still image mockup as a visual reference to direct the edit rather than describing changes in text. Additional capabilities: swap products for new colorways, change backgrounds for seasonal campaigns, fix lighting, remove distractions, restyle to a different aesthetic. Preview any proposed edit as a still image before committing to video generation — significantly reducing wasted compute credits. Output: 30 seconds at 1080p. Available same-day on all paid Runway plans. Original Gen-4 features include reference-guided character consistency and flexible 2–10 second generation durations.

One of the highest-rated AI video platforms in 2026. Kling 3.0 (February 3-5, 2026) scored 8.1/10 from Curious Refuge — 'the highest-scoring AI video model we've reviewed to date' — and outperforms both Sora 2 and Google Veo 3.1 in head-to-head testing. The Kling 3.0 breakthrough: multi-shot generation of up to 6 camera cuts in a single generation, each with customizable duration and camera movement. Previously this required generating each shot separately and stitching in editing. Also features native 4K at 60fps and lip sync in 5 languages. Kling 2.6 added native simultaneous audio + video in one pass.