AI&Design
All change articles

Runway · Announcement · Official source

Runway opens Gen-4.5 text-to-video to all paid plans

Runway introduced Gen-4.5 on December 1, 2025, with a staged rollout. On December 11, its app changelog recorded availability for all paid plans. The December workflow generated a video directly from a written description of the scene, action and camera. A designed opening image became an input to Gen-4.5 in the separately dated January 21 release. December research announcement · December announcement video · Runway changelog

What users discussed

In the December 1 announcement thread, Runway's account told readers to expect gradual access and later explained that rollout varied even within the Pro tier. By December 11, a prospective subscriber was asking whether Gen-4.5 ran in Unlimited's Explore Mode. The announcement account replied that it did, and another user reported trying a text-to-video generation there. The changelog establishes paid-plan availability. The discussion records how participants understood one plan's mode. December 1 rollout thread · December 11 Unlimited question · Runway changelog

A separate December 11 thread asked for hands-on impressions. One commenter described promising action alongside weaker-looking people and suggested replacing subjects in a separate Aleph pass. A different creator posted a clip labeled as a 720p Gen-4.5 test and confirmed in the thread that it was made from text. These public posts illustrate specific uses and questions within the December 4–25 collection window. The December 1 thread supplies the rollout context. Experience thread · Creator's 720p test

What the examples show

In a December 22 DataCamp walkthrough, François Aubry published written prompts and embedded outputs for a night market and a grocery checkout. He reported that the crowded market captured much of the description, but a background scooter and some people disappeared. In the grocery clip, he reported a missing line of customers and unnatural item movement. His two seesaw tests produced different reactions to an elephant and a mouse, though he reported that the elephant stepped onto the seesaw instead of falling as requested. The examples document enough detail to inspect prompt interpretation and continuity during short-shot exploration. The motion observations come from Aubry's write-up, with playback unverified in this review. DataCamp first-person tests

A December 17 post from Runway's announcement account shared a yarn stop-motion sample and a written recipe for crochet forms, soft lighting and visible fiber. The author confirmed straight text-to-video in the replies. For a designer, these examples suggest that a written brief can get a visual treatment or camera idea onto the screen quickly enough to review it. The published examples leave retry costs, repeatability and production adoption unmeasured. Yarn stop-motion demonstration · DataCamp first-person tests