Recraft V4 Vector

Overview

This model takes text prompts and outputs native SVG files containing actual vector paths and structured layers. Unlike raster image generators, it produces scalable graphics that retain crisp edges at any resolution. On Kyma, requests route through an OpenAI-compatible endpoint with automatic failover if a serving path degrades. The API returns exact request costs in the usage object, identifies the active model via the X-Kyma-Model header, and applies prompt caching to repeated prefixes at this model’s cached input rate. It accepts text input only and does not support vision, reasoning, or structured JSON outputs. Generation speed is medium, and it is optimized for graphic assets rather than photorealistic imagery or complex scene composition.

Specs

Field	Value
Model ID	`recraft-v4-vector`
Best for	Logos, icons, illustrations as editable SVG
Context window	0
Input modalities	Text
Output modalities	Image
Tool calling	Yes
Structured outputs	Yes
Prompt caching	Yes
Speed	Medium
Cost band	Balanced
Release stage	Stable

Pricing

	Per 1M tokens
Input	$0.20
Output	$0.60

Use this when

Brand Logo Generation — Create scalable logos with clean paths that designers can refine in vector editors.
UI Icon Sets — Generate consistent, resolution-independent icons for web and mobile interfaces.
Editable Vector Illustrations — Produce layered graphics ready for direct manipulation in Figma or Illustrator.
Technical Diagram Assets — Output precise, scalable shapes and lines for documentation and presentations.

Not ideal for

Do not use this model for photorealistic rendering, complex scene generation, or tasks requiring image-to-image editing, as it outputs structured vector paths rather than raster pixels.

Pick something else when

You need photorealistic images or complex scenes: use flux-2-pro or imagen-4-ultra.
You need fast raster generation for prototyping: use imagen-4-fast or flux-1.1-ultra.
You need high-fidelity raster illustrations: use recraft-v4 or ideogram-v3.

Example

from openai import OpenAI

client = OpenAI(base_url="https://kymaapi.com/v1", api_key="ky-...")

response = client.embeddings.create(
    model="recraft-v4-vector",
    input=["first document", "second document"],
)

Agent query example

Ask the API which models fit, instead of hardcoding an id:

curl "https://kymaapi.com/v1/models?recommended_for=generate-images&tools=true&quality_tier=frontier-open"

FAQ

Does this model output raster or vector files? It outputs native SVG files with actual paths and layers, not PNG or JPEG raster images. Can I edit the generated files in design software? Yes, the SVG output is structured for direct import and editing in Figma, Illustrator, and other vector tools. Does Kyma support prompt caching for this model? Yes, repeated prompt prefixes are billed at this model’s cached input rate, reducing costs for iterative design workflows.

​Overview

​Specs

​Pricing

​Use this when

​Not ideal for

​Pick something else when

​Example

​Agent query example

​FAQ