Runway Gen-4 guide: 5/10-second image-to-video prompts

Runbo Li
Runbo Li
·
· 3 min read
Runway

Quick answer

Runway Gen-4 is still available, but Runway now labels it an older image-to-video model and recommends Gen-4.5 for current text-to-video and image-to-video work. Use Gen-4 Turbo to test a required source image and motion prompt at 5 credits per second; Gen-4 costs 12 credits per second. Both make 5- or 10-second clips. Use Gen-4.5 when you need the current model, text-only input, or a duration from 2 to 10 seconds.

Animate a source image

Start with the same source image and one simple motion prompt. Inspect the full clip before adding more actions or spending credits on a longer generation.

Try Image-to-Video

Runway Gen-4 status and specs

Runway model

Current role

Inputs

Duration

Cost

Output

Gen-4

Older image-to-video model

Required image + motion prompt

5 or 10 seconds

12 credits/second

720p-class aspect-ratio presets; 24 fps

Gen-4 Turbo

Lower-cost iteration on the older workflow

Required image + motion prompt

5 or 10 seconds

5 credits/second

Same listed aspect-ratio presets; 24 fps

Gen-4.5

Runway's current text-to-video and image-to-video starting point

Text, or text + image

2 to 10 seconds

12 credits/second

720p; 24 or 25 fps

Facts were checked September 13, 2026 against Runway's Gen-4 guide, Gen-4.5 guide, getting-started guide, plan pricing and API pricing. Runway explicitly marks Gen-4 as an older generation. This page preserves Gen-4 search intent while separating it from the current default.

How to use Runway Gen-4

1. Choose Gen-4 or Gen-4 Turbo

In Runway Tool mode, open Video and select Gen-4 or Gen-4 Turbo. Both require an input image. Choose Turbo for lower-cost iterations; switch only when a representative result gives you a reason to spend the higher rate.

2. Upload the first frame

Use a clean image at the composition and aspect ratio you want. The image supplies the subject, lighting, style and starting composition. Cropping to a different output ratio can remove important details before motion is generated.

3. Describe motion, not the whole picture

Runway's official Gen-4 prompting guide recommends a simple, positive description of subject motion, camera motion, scene motion and broad style. Do not repeat every visible object or use negative instructions.

[subject motion]. [camera motion]. [scene response]. [motion style].

The woman turns toward the window and smiles. The locked camera remains still. Curtains move in a light breeze. Natural live action.

Workflow recipe output preview: Body Twist

Workflow recipe

A cinematic fashion-style motion scene of a young woman standing naturally. The camera starts with a soft medium shot, slight push-in, clean studio lighting. The woman slowly turns her upper body and head toward the camera in a smooth, natural motion, her long hair flowing gently with the movement. She then raises both hands up behind her head in a relaxed, confident pose. As soft rhythmic music plays, she begins to sway her hips gently from side to side in a natural, fluid motion, subtle and elegant, not exaggerated. Her facial expression becomes softer and more confident, with a slight smile. Her clothing and hair move naturally with her body motion, realistic fabric physics. Camera remains steady with a slight cinematic orbit to enhance depth. Lighting stays soft and flattering, highlighting skin tones and texture. Style: cinematic, fashion editorial, natural movement, smooth animation, high detail. Negative: no jitter, no stiff movement, no distortion, no unnatural bending, no flickering.

Model
kling-3.0
Resolution
720p
Duration
8 seconds
Frame rate
24 fps

The camera slowly dollies toward the product. Reflections move across the metal surface. Controlled studio motion.

The subject walks toward the left edge of frame. A handheld camera tracks beside them. Dust trails behind each step.

4. Pick 5 or 10 seconds

Gen-4 and Turbo offer 5- and 10-second generations. Five seconds costs 60 credits on Gen-4 or 25 credits on Turbo; ten seconds costs 120 or 50 credits. Treat one generation as one shot. A transformation, camera orbit and scene change in the same short prompt can compete with each other.

5. Generate, inspect and change one variable

Watch the entire result for subject identity, hands, object continuity, camera direction and the final frame. If the motion is wrong, simplify the prompt before changing the source image. If the composition is wrong, fix the image before adding prompt detail. A fixed seed can help explore similar motion, but it does not make two outputs identical.

6. Continue or export

Runway's current Gen-4 interface can retime, expand, reuse the current frame, upscale or pass the result into other workflows. Keep a clean master and verify the delivered aspect ratio and rights for every source asset.

When to use Gen-4.5 instead

Gen-4.5 is Runway's current starting point for text-to-video and image-to-video. It supports 2-to-10-second durations, costs 12 credits per second and outputs 720p at 24 or 25 fps. Use it when you need text-only generation, variable duration, current prompting guidance or the latest Runway model path.

For a complete subscription budget, use the current Runway pricing guide. For different platforms and workflows, see Runway alternatives. Compare model costs using the same duration and accepted-output rule; a cheaper attempt can cost more if it needs more retries.

Runway Gen-4 vs Magic Hour

Runway Gen-4 is a specific older image-to-video model inside Runway. Magic Hour is a broader creation platform. Use Magic Hour's Image-to-Video for source-image animation, Text-to-Video for prompt-only shots, or Video-to-Video to transform existing footage. Compare the exact model, duration, resolution, controls and accepted-output cost for your job rather than treating either product name as one model.

Frequently asked questions

Yes. Runway's current help center still lists Gen-4 and Gen-4 Turbo in Tool mode, but labels the guide as covering older models and directs users to Gen-4.5 for current text-to-video and image-to-video capabilities.

No. Gen-4 and Gen-4 Turbo require an input image plus a text motion prompt. Gen-4.5 supports both text-to-video and image-to-video.

Gen-4 costs 12 credits per second: 60 credits for five seconds or 120 for ten. Gen-4 Turbo costs 5 credits per second: 25 or 50 credits. Runway Dev sells API credits at $0.01 each, so API Gen-4 Turbo is $0.25 for five seconds or $0.50 for ten; consumer subscription credits are bundled differently.

Runway lists aspect-ratio presets including 1280×720 landscape, 720×1280 vertical and 960×960 square at 24 fps. Its interface also offers a separate 4K upscale action; that does not change the model's listed generation resolution.

Describe the desired motion in direct, positive language. Start with one subject action and one camera behavior. Add scene response or style only after the basic motion works.

For the older Veo 3 and Runway Gen-4 comparison and its current-status limits, see Veo 3 vs Runway Gen-4.

Runbo Li
Runbo Li
CEO of Magic Hour
Runbo Li is the Co-founder and CEO of Magic Hour, where he builds AI video and image tools for content creation. He is a Y Combinator W24 founder and former Data Scientist at Meta, where he worked on 0-1 consumer social products in New Product Experimentation. He writes about AI video generation, AI image creation, creative workflows, and creator tools.
View author →

Continue Reading

Runway
How to use Runway: current guide for creators
veo3
Google Veo 3.1: a beginner's guide to AI video
seedance
Seedance 1.0 guide: release, capabilities and current status
Flux <> MH
FLUX.1 Kontext tutorial: edit images, prompts & FLUX.2
Handmade collage showing a portrait becoming an animated character video with audio
Hedra AI: Character 3 tutorial, pricing & Omnia comparison
pika
How to use Pika AI (2026): video, PikaSwaps, audio & API