Runway AI Timeline: Model Releases & Version History

Runway began as a way for artists to use machine-learning models, then expanded into AI video generation, image creation, and video editing. Its later releases brought performance-driven character animation, conversational production tools, and interactive world models.

The timeline records major model announcements, product launches, API releases and retirements, tracing Runway’s development back to its 2018 prototype.

Last updated: October 28, 2026.

2026

Date	Milestone
October 27, 2026	Runway Ads. General availability across paid plans. Connect ad accounts, generate creative variants and publish approved ads.
October 26, 2026	App Studio. Released for paid plans. Build and share custom creative apps from plain-language descriptions.
September 30, 2026	Runway Ads announcement. Announced as a product for generating, publishing and iterating on advertising creative.
September 30, 2026	Praxis-1. Announced for robotics. A world action model enters partner testing, with public weights planned for a later release.
September 23, 2026	DaVinci Resolve Plugin. Integration released. Generate and edit with Runway from inside DaVinci Resolve.
September 8, 2026	Runway Adobe Plugins. Integrations released for Premiere Pro and After Effects, with image/video generation and Aleph editing inside the editors.
September 3, 2026	GWM Worlds 2. Research preview adds real-time audio-video world simulation with text actions and camera control.
September 2, 2026	Runway Dev MCP. Released. Coding agents can inspect model schemas, configure integrations and debug Runway Dev tasks.
August 31, 2026	Solaris. Interface world model introduced. Generate interactive visual interfaces frame by frame in response to user actions.
August 20, 2026	Runway Ruby. Released. An SDR-to-HDR color-grading model for existing video, available to Max+ users in Tool Mode and through Runway Dev.
July 30, 2026	Video Editor Projects retirement. The legacy video project editor is retired. Timeline Studio replaces Video Editor Projects.
July 30, 2026	Gen-3 Alpha Turbo / Gen-4 Aleph. API retirement. The `gen3a_turbo` and `gen4_aleph` models are removed from Runway Dev.
July 23, 2026	Runway Dev. Developer platform launched with model access, Media Router, Recipes, Workflows and Characters.
July 21, 2026	Model Routers. API availability. Route generation requests across models using cost, latency and quality preferences.
July 9, 2026	Timeline in Agent. Studio’s timeline editor moves into Agent for clip assembly, refinement and export within a conversation.
June 25, 2026	Agent 2.0. Released to all users. Adds campaign creation and iteration based on supplied marketing performance data.
June 18, 2026	Studio. Released to all plans. Trim, combine and reorder clips, then export an assembled video.
June 2, 2026	Aleph 2.0. API release. Edit videos with text prompts and optional images assigned to specific timestamps.
May 27, 2026	Runway MCP. Hosted connector released. Use a Runway account to generate media from compatible agents.
May 21, 2026	Aleph 2.0 & Edit Studio. Paid web release. Apply text or edited-frame instructions to videos up to 30 seconds, including edits across multiple shots.
May 13, 2026	Runway Agent. Released. A conversational workspace for planning and producing finished videos with multiple scenes and audio.
April 10, 2026	All Tools page removal. The All Tools page is removed. Legacy Magic Tools move to newer Apps and Tool Mode experiences.
March 9, 2026	Runway Characters. API and web demo launch. Create real-time conversational video characters from a reference image, powered by GWM-1.
February 20, 2026	Third-party models in Runway. Paid-plan rollout expands image and video generation options inside Runway, including models from Kling, WAN and OpenAI.
February 10, 2026	Gen-4.5. API release with text-to-video and image-to-video generation.
January 21, 2026	Gen-4.5 Image to Video. Image-to-video released for paid plans, with an image supplied as the opening frame.

2025

Date	Milestone
December 11, 2025	Gen-4.5. Available to all paid plans following the December 1 announcement.
December 11, 2025	GWM-1. World-model family announced. GWM Worlds, GWM Avatars and GWM Robotics use action-conditioned, real-time generation.
December 1, 2025	Gen-4.5. Announced with improvements in motion quality, prompt adherence and visual fidelity.
October 24, 2025	Workflows. Released for paid plans. Connect models and intermediate processing steps in a node-based workflow.
October 14, 2025	Apps. Released on the web. Task-specific creative workflows become available through dedicated apps.
September 23, 2025	Third-party models in Runway API. The API adds Google Veo 3 and Nano Banana alongside Runway models.
August 21, 2025	Game Worlds Beta. Public beta launches with interactive stories, generated images and custom game worlds.
August 19, 2025	Gen-4 Image Turbo. Released on the web and API. A faster Gen-4 image model also becomes available to free-plan users.
August 1, 2025	Aleph. API release for instruction-based video editing and transformation.
July 31, 2025	Aleph. Available to all paid plans. Edit existing video through instructions that change objects, viewpoints, lighting or style.
July 25, 2025	Aleph. Introduced as an in-context video model for editing and transforming existing footage.
July 21, 2025	Act-Two. API release for performance-driven character animation.
July 17, 2025	Act-Two. Released for paid plans. Performance transfer expands to head, face, body and hand motion.
June 13, 2025	Runway MCP Server. Released on GitHub for image and video generation through compatible assistants.
June 12, 2025	Chat Mode. Released for all plans. Generate and refine Gen-4 images, videos and References in a conversational workspace.
May 16, 2025	Gen-4 Image in Runway API. API release. Generate images with Runway’s Gen-4 Image model.
May 11, 2025	Gen-2 retirement. Gen-2 fully retired after access reductions began in March. Motion Brush is retired with the model.
May 7, 2025	Gen-4 Images and References for Free Plan. Gen-4 image generation and References become available to free-plan users.
April 30, 2025	Gen-4 References. Released for paid plans. Use reference images to preserve characters, objects and locations across new images.
April 7, 2025	Gen-4 Turbo. Released to all users as a faster, more efficient Gen-4 video model.
April 6, 2025	Gen-1 retirement. Gen-1 retired as a standalone tool.
April 1, 2025	Gen-4 video. Video generation becomes available to paid users after the March 31 introduction.
March 31, 2025	Gen-4. Introduced with reference-based consistency for characters, objects and locations across images and video.
January 17, 2025	Frames. Available on Unlimited and Enterprise plans following its November 2024 introduction.

2024

Date	Milestone
November 25, 2024	Frames. Introduced as an image-generation model focused on stylistic consistency and control.
November 1, 2024	Camera Control in Gen-3 Alpha Turbo. Camera Control released. Set the direction and intensity of camera movement in generated scenes.
October 22, 2024	Act-One. Release announced with a gradual rollout. Transfer an actor’s performance to a character inside Gen-3 Alpha.
September 16, 2024	Runway API. API launch with Gen-3 Alpha Turbo. Initial access is limited to selected partners, with wider rollout planned.
September 13, 2024	Video to Video in Gen-3 Alpha. Video-to-video released for paid plans. An input clip guides movement while prompts change the visual treatment.
August 14, 2024	Gen-3 Alpha Turbo. Released. A faster, lower-cost version of Gen-3 Alpha.
July 29, 2024	Gen-3 Alpha Image to Video. Image-to-video released for paid plans, with an image used as the first frame.
July 1, 2024	Gen-3 Alpha Text to Video. Text-to-video becomes available on paid plans following the June 17 announcement.
June 17, 2024	Gen-3 Alpha. Announced. A new video model trained on multimodal infrastructure, with improved motion, consistency and fidelity.
March 22, 2024	Lip Sync. Released. Animate a face to match the speech in an audio track.
January 8, 2024	Custom Voices. Released for Pro and higher plans. Create a voice from a short audio sample for speech and lip-sync work.

2023

Date	Milestone
December 11, 2023	General World Models. Research program announced to develop models that represent environments and simulate their dynamics.
November 23, 2023	Motion Brush. Released for Gen-2 image-to-video. Paint an area and specify how it should move.
June 7, 2023	Gen-2 web and iOS release. Gen-2 opens on the web and iOS after its restricted-access phase.
March 27, 2023	Gen-1 web release. Gen-1 becomes available in the Runway web app.
March 20, 2023	Gen-2. Announced. Generate video from text, with access initially staged through a waitlist.
February 6, 2023	Gen-1. Research published. Gen-1 transforms source footage through text or image guidance while preserving its structure.

2022

Date	Milestone
October 20, 2022	Stable Diffusion v1.5. Runway publishes the Stable Diffusion v1.5 checkpoints used by its text-to-image and image-to-image tools.
August 22, 2022	Stable Diffusion. Stable Diffusion co-released by Stability AI, LMU Munich researchers and Runway ML.

2021

Date	Milestone
December 20, 2021	Latent Diffusion Models. Research published. Work by LMU Munich, Heidelberg University and Runway introduces image generation in a compressed latent space.
July 20, 2021	Sequel video editor. Sequel released. Runway adds a machine-learning-powered video editor.

2020

Date	Milestone
November 17, 2020	Green Screen. General availability after its November 11 introduction. Green Screen uses machine learning to cut subjects out of videos.
August 5, 2020	Runway Web. Runway Web introduced following a July 23 soft launch.
April 28, 2020	Hosted Models. Hosted Models released. Publish a model at a hosted URL for access outside the desktop app.

2019

Date	Milestone
June 4, 2019	Runway public beta. Public beta released. Creative users can run machine-learning models through the Runway application.

2018

Date	Milestone
April 9, 2018	Runway prototype. Cristóbal Valenzuela demonstrates the early Runway project with OpenPose and Processing and invites beta testers.

How Runway’s model families fit together

The Gen series spans several kinds of video generation. Gen-1 starts with existing footage, while Gen-2 introduced text-driven video creation. Later Gen models added stronger control over motion and scene consistency. Frames belongs to the image-generation branch and is now named Gen-4 Images. References supplies visual guidance for recurring subjects and settings.

Aleph focuses on changing footage that already exists. Act-One and Act-Two transfer a recorded performance to a character. These tools address different production tasks, so their version numbers do not form one shared sequence.

GWM-1 extends video modeling into interactive simulation. Built on Gen-4.5, its autoregressive models generate successive frames conditioned on actions such as camera movement or robot commands. This research branch supports environments that respond during generation, beyond the fixed clips used in conventional video production.