Chinese short-video company Kuaishou launches Kling Video O1, saying it is the first multimodal AI model to unify video generation, editing, and post-production
Brand-New Creative Engine for Endless Possibilities! Input anything. Understand everything. Generate any vision. With true multimodal understanding, Kling O1 unifies your input across texts, images, and videos — making [video] @ehuanglu : finally.. nano banana pro for video is here Kling O1 just dropped on Higgsfield, you can edit video with text and image reference.. replace actor, change bg&lighting, add vfx and most importantly.. it keeps the rest of the scene consistent here's how to use it: [video] @kling_ai : With Kling O1, you can be ANYWHERE you want! Whether it's a change in background, environment, or even a weather, simply have Kling O1 put you in your desired world seamlessly. [video] @mayorkingai : Nano Banana for video is here. Kling just dropped o1. Invideo just turned it into a full VFX house. AI video is finally something you can control - not just generate. Here are the insane things you can do with it 👇 PS: You've got a 7-day window where Kling o1 runs free & [video] @invideoofficial : This is VFX House - a full VFX studio built on @Kling_ai o1 inside invideo. AI Companies have been LYING to creators for 2 years. Cool demos but - Zero control. Zero reliability. Zero continuity. Today... that ends. The chaos era is over. The control era begins now. Kling o1 [video] @esotericcofe : it's been a while since i've been genuinely impressed by a video model, but kling o1 can literally just remove objects in a video. no masking needed. prompt: “remove the tram” [video] See also Mediagazer
Context & Ripple Effects
Kling’s earlier product work emphasized faster, cheaper clip creation, including a claimed 1080p output in under a minute in Kuaishou’s prior Kling update. O1 expands that pitch from generating a clip to altering and finishing one from combined text, image, and video inputs.
That workflow framing matters because Kling had already been positioned as a potential revenue driver for Kuaishou in the company’s earlier revenue outlook. The launch tests whether video AI can become a repeat-use production tool rather than a prompt-only feature.
First-order effects
- Kling users can use one model for generation, scene changes, and post-production-style modifications, potentially reducing handoffs between separate creative tools for supported tasks.
- Kuaishou gains a more comprehensive product claim with which to sell Kling to creators and production-oriented users; the company’s share-price decline shows investors did not immediately treat the announcement as sufficient proof of commercial impact.
Second-order effects
- Rival video-model providers, including those pursuing text-to-video products such as Meta’s Movie Gen, face pressure to match editing control and cross-media inputs rather than compete chiefly on generated clips.
- VFX teams and creative-tool vendors may see demand shift toward tools that preserve consistency while making targeted changes, putting a premium on controllable editing workflows over standalone generation.
Third-order effects
- If integrated models reliably handle generation and revision in the same workflow, video AI competition will increasingly center on production-system fit, not just output quality or clip length.
- That shift could concentrate value in platforms that own user inputs, editing context, and distribution, while making provenance and control mechanisms more important as synthetic footage moves deeper into production.
The trend: Generative-video products are evolving from single-prompt creation tools into workflow-native systems that ingest existing media and support iterative production.