Models
Video generation models
By KrafLayer team6 min read2026-04-06
TL;DR
Compare AI video models by input type, duration, output controls, and workflow features.
Text-to-video vs. image-to-video
Text-to-video models generate motion from a prompt alone. Image-to-video models animate from a starting frame, giving you direct control over the first shot.
Use the table to compare duration range, supported inputs, and available controls.
How to read the table
- Input tells you whether the model starts from prompt only or from prompt plus a start frame.
- Output summarizes duration, resolution, and framing controls.
- Features call out start-frame input, end-frame guidance, audio, and other workflow-changing options.
- Use Costs by model and task when you need the current credit cost per generated second.
Related KrafLayer docs
- [Costs by model and task](/docs/generation-cost)
- [Plans and price](/docs/pricing-and-credits)
- [AI image-to-video guide](/docs/ai-image-to-video-guide)
- [Sora series](/docs/sora-models)
- [Kling series](/docs/kling-models)
Related KrafLayer tools
- AI product image tools — Browse the full tool list for ecommerce image editing and product visual workflows.
- Listing main and detail images — Generate ecommerce listing main images and detail-page product visuals from product references.
- AI scene compose — Place products into controlled commercial scenes without losing product clarity.
- AI background replacer — Move a product into a cleaner studio, lifestyle, or campaign background.
- AI reference image editor — Use extra references to guide product identity, material, style, or composition changes.
- Reference-style product images — Generate ecommerce product images from competitor, brand, or campaign reference styles while preserving your own product identity.