Runway Gen-4 Image Technical Specifications
| Item | Verified information |
|---|---|
| Model name | Runway Gen-4 Image |
| API model ID | gen4_image |
| Provider | Runway |
| Modality | Text and reference-image input; generated image output |
| Core task | Text-to-image and image-to-image generation |
What Is Runway Gen-4 Image Generation?
Gen-4 Image is Runway's image-generation model for creating still visuals from prompts and reference images. It is positioned for controllable, consistent image creation rather than a general language-model workflow.
Features and Highlights of Runway Gen-4 Image
- Generates new scenes from natural-language prompts.
- Uses reference images to guide subject, composition, or visual style.
- Supports iterative art direction through prompt and reference changes.
- Useful for concept art, storyboards, product imagery, and previsualization.
- Can provide source frames for downstream Runway video workflows.
Runway Gen-4 Image Benchmark Performance
Runway has not published a standardized public benchmark table for the gen4_image API ID in the sources below. Quality is best evaluated with a task-specific image set.
Runway Gen-4 Image vs Gen-4 Image Turbo vs Gen-4 Turbo
| Model | Best fit | Publicly verified distinction |
|---|---|---|
| Gen-4 Image | Highest-control still-image work | Image-generation model |
| Gen-4 Image Turbo | Faster image iteration | Turbo variant; exact limits are deployment-specific |
| Gen-4 Turbo | Fast video generation workflows | Video-focused Turbo variant |
Runway Gen-4 Image Limitations
Runway does not publish a universal context window or max-output-token value for this visual model. Results can vary with prompt specificity, reference quality, and safety filtering.
Runway Gen-4 Image Use Cases
Use gen4_image through CometAPI for campaign key art, game or film concept frames, product mockups, storyboard panels, and reference-guided visual exploration. Keep prompts explicit about subject, camera, lighting, and aspect ratio, then select accepted images for later video generation.