Unofficial community guide for Runway Gen-4. Not affiliated with Runway. All trademarks belong to their owners.
This is a working guide to runway gen 4, Runway's image-to-video model, written from the Runway help center article "Creating with Gen-4 Video", the research post introducing the model, and the Morphic model page that hosts Gen-4 Image. It covers what the model takes as input, how credits are charged, which settings exist, and where people get stuck. Anything not stated on those pages is left out rather than guessed.
Gen-4 needs an input image for every clip. If you want to start from text as well, try Kyncept - Veo 3.1 text-to-video and image-to-video in the browser. It also generates the still images you would otherwise need to bring.
According to Runway's help center, Gen-4 creates fast, controllable and flexible video generation that can sit beside live action, animated and VFX content. It produces clips of 5 or 10 seconds from an input image plus a text prompt you provide; the image is required, not optional. Output is 24 fps at fixed resolutions per aspect ratio: 1280x720 (16:9), 720x1280 (9:16), 960x960 (1:1), 1104x832 (4:3), 832x1104 (3:4) and 1584x672 (21:9). Text prompts are capped at 1000 characters. It is available on the web app and the iOS app.
There are two variants. Gen-4 is the full model at 12 credits per second of video; Gen-4 Turbo is the faster, cheaper version at 5 credits per second, which Runway describes as a strong model for quickly iterating ideas. Runway's own recommendation is to test in Turbo and switch to Gen-4 as needed. Note that the help center now flags this article as covering an older generation of Runway models and points to Gen-4.5 for text-to-video and image-to-video, the Animate Frames app for keyframes, and Edit Studio Aleph 2.0 for video-to-video. Gen-4 is still selectable, but it is no longer the newest option.
- Sign in to the Runway web app and open Tool mode.
- Select Tool in the left-hand sidebar, then the Video tab at the top of the generation panel.
- Open the model dropdown at the bottom of the panel and pick Gen-4 or Gen-4 Turbo.
- Upload the input image. Gen-4 requires one; there is no text-only path.
- Write a prompt of up to 1000 characters describing the motion you want.
- Choose 5 or 10 seconds and the aspect ratio, then generate. Iterate in Turbo first, then rerun the keeper in Gen-4.
| Gen-4 | Gen-4 Turbo | |
|---|---|---|
| Cost | 12 credits per second | 5 credits per second |
| 5-second clip | 60 credits | 25 credits |
| 10-second clip | 120 credits | 50 credits |
| Platform availability | Web, iOS app | Web, iOS app |
| Base prompt inputs | Text, image (required) | Text, image (required) |
| Prompt limit | 1000 characters | 1000 characters |
| Frame rate | 24 fps | 24 fps |
The help article prices everything in credits, not currency. How many credits your plan includes, and what they cost, is on Runway's pricing page, so check there for current numbers.
- Budget by the second. A 10-second Gen-4 clip is 120 credits; the same clip in Turbo is 50. Ten Turbo drafts cost less than five Gen-4 drafts, which is why Runway suggests drafting in Turbo.
- The image does most of the work. Because the model is image-to-video, composition, lighting and subject identity come from the still. Fix the image before spending credits on prompts.
- Pick the aspect ratio for the destination. Each ratio has one fixed output size; there is no 1080p option listed for Gen-4 in the help article. Morphic's page for Gen-4 Image lists 1080p output on its platform, so check the platform you are actually using.
- Prompt for motion, not for the scene. The Morphic examples are short and motion-focused: "gentle breeze moving the hair, soft blink, subtle smile forming, natural breathing movement" or "slowly rotate this product with smooth studio turntable motion". You have 1000 characters, but you rarely need them.
- Check the newer models first. The help center itself says Gen-4.5 covers text-to-video and image-to-video now. If you are starting a new project, compare before committing a workflow to Gen-4.
- Video-to-video is a different tool. Gen-4 is image-to-video only. For editing existing footage the help center points to Edit Studio Aleph 2.0.
| Runway Gen-4 | Runway Gen-4 Turbo | Gen-4 Image on Morphic | Kyncept | |
|---|---|---|---|---|
| Input | Image (required) plus text | Image (required) plus text | Image | Text or image |
| Clip length | 5 or 10 seconds | 5 or 10 seconds | Up to 10 seconds | Not covered here |
| Output | Up to 1280x720 or 1584x672, 24 fps | Same as Gen-4 | 1080p, 24 fps | Not covered here |
| Cost | 12 credits per second | 5 credits per second | See Morphic pricing | See the site |
| Where | Runway web and iOS | Runway web and iOS | Morphic Studio | Browser |
| Also generates images | No | No | No | Yes |
Can Gen-4 make a video from text alone? No. The help article lists the image as a required input. Runway points to Gen-4.5 for text-to-video.
How long can a clip be? 5 or 10 seconds. Longer sequences mean chaining separate generations.
What is the difference between Gen-4 and Gen-4 Turbo? Turbo is faster and costs 5 credits per second instead of 12. Runway recommends drafting in Turbo and moving to Gen-4 when Turbo results fall short.
Is there an API? The pages this guide is based on describe the web and iOS apps only. Check Runway's developer documentation for API availability.
Where can I use Gen-4 outside Runway? Morphic lists Runway Gen-4 Image on its platform as an image-to-video model with 1080p output and up to 10 seconds per generation.
Gen-4 is a solid image-to-video model with a clear credit table, and the Turbo-then-Gen-4 loop is a sensible way to spend credits. Its constraint is the required input image: every clip starts with a still you have to source or make somewhere else. Try Kyncept - Veo 3.1 text-to-video and image-to-video in the browser if you want both paths in one place. Kyncept generates video from a text prompt or from an image with Veo 3.1, and it generates images too, so the still you would have brought to Gen-4 can be made in the same tab. For teams that go from idea to clip without a separate image pipeline, that is the shorter route.
Last reviewed: 2026-09-22