Published: October 1, 2026
Image Model
Recraft V4.1 is the newer generation of Recraft's design-first model, following base V4 with quieter, more natural photorealism and sharper illustration and vector work. It reads short prompts the way a creative director would, filling gaps with its own taste instead of defaulting to generic stock-photo compositions. It ships on Toolkit as two separate models: Recraft 4.1 Image for raster image generation, and Recraft 4.1 Vector for structured, editable SVG output. Each has a simple Resolution setting (1K or 2K) and an Aspect Ratio picker.
Variants at a Glance
Variant |
Type |
Modalities |
Resolution |
Talking Points |
|---|---|---|---|---|
| Variant | ||||
| Recraft 4.1 Image | Image | Text-to-Image | 1K or 2K (user-selected) | Sharper prompt control and cleaner composition than base V4; built for brand and editorial work |
| Recraft 4.1 Vector | Image | Text-to-Vector | 1K or 2K (user-selected) | Structured, editable SVGs for logos, icons, and illustration systems |
Key Features
Image Generation
- Text-to-Image: Quieter, more natural photorealism and cleaner composition than base V4; choose 1K or 2K resolution per generation
Vector Generation
- Text-to-Vector: Structured, editable SVGs for logos, icons, and illustration systems — a separate Toolkit model from Image, not a shared mode within it; choose 1K or 2K resolution per generation
Technical Capabilities
Spec |
Details |
|---|---|
| Modalities | Text-to-Image (Recraft 4.1 Image model), Text-to-Vector (Recraft 4.1 Vector model) — two separate Toolkit models, not a shared mode on one model |
| Resolution | 1K or 2K, user-selected at generation time on both models |
| Aspect Ratios | 16:9, 9:16, 4:3, 3:4, 1:1 |
| Output Formats | WEBP (raster, Image model), SVG (vector, Vector model) |
Limitations
- No single model outputs both raster and vector — Image and Vector are separate Toolkit models, so pick the one that matches the output format you need.
- Does not support V4's user-created Styles feature.
Prompting Tips
- Keep prompts short on V4.1 — the model fills gaps with its own art direction rather than defaulting to generic stock-photo compositions.
- Use 1K for faster, lower-cost iteration; switch to 2K when you need higher resolution for hero imagery, campaign work, or print.