Midjourney debuts first AI video generation

Midjourney, best known for its AI‐driven image creation tools, this week unveiled V1, its inaugural video generation model that animates stills into short clips.

Available now to its roughly 21 million users via the Midjourney website and Discord, V1 marks the company’s strategic pivot from static imagery toward full multimedia content generation.

V1 operates as an image-to-video system: users upload or select an existing Midjourney image and receive four distinct, five-second video renditions by default.

The model can then extend a clip by up to 16 additional seconds, in four-second increments, culminating in a maximum length of 21 seconds.

Access to V1 is bundled into Midjourney’s subscription tiers, with entry-level plans priced at $10 per month.

Customization is a core feature, with V1 offering “automatic” mode for hands-off generation or a “prompted” workflow where users guide motion through text inputs.

Other customisation settings allow users to fine-tune “motion intensity”.

Meanwhile, the platform's image gallery interface now includes an “animate” button, providing another option to create, powering transition from still to video.

It comes amid ratcheting competition among multimedia AI, weeks after Google wowed with its Veo video editor and a couple of months after OpenAI upgraded pictures and video in ChatGPT and Sora.

Midjourney meanwhile faces copyright friction, with suits coming from Disney and Universal Studios alleging unauthorized use of film footage in AI training data.

Nevertheless, Midjourney as the early groundbreaker in AI art and image creator is still keeping up with its deep-pocketed Silicon Valley rivals.