# Arrow 1.1

Arrow 1.1 is QuiverAI's default SVG generation model, balancing speed and quality. Arrow 1.1 turns text prompts into editable vector graphics and vectorizes existing images into clean SVG files.

- **Model ID:** `quiverai/arrow-1.1`
- **Type:** image
- **Providers:** quiverai
- **Context window:** 131,072
- **Maximum output tokens:** 131,072
- **Pricing:** See the provider details below
- **Canonical page:** https://vercel.com/ai-gateway/models/quiverai/arrow-1.1

## Supported parameters

Detailed capability metadata has not been reported for this model.

## Example

```ts
import { experimental_generateImage as generateImage } from 'ai';

const result = await generateImage({
  model: 'quiverai/arrow-1.1',
  prompt: 'A red balloon on a wooden table.'
});
```

## About

Arrow 1.1 is an SVG generation model from QuiverAI, an AI lab focused on vector design. Most image models produce pixels. Arrow 1.1 writes vector graphics directly as structured code, making designer-like choices about control points, color, and gradients instead of approximating shapes from a raster image.

Arrow 1.1 improves on the previous Arrow generation in design quality, prompt following, and SVG precision, especially in finer details. Output favors primitives such as shapes, text, and basic geometry over stacked paths. The resulting files are easier to edit, inspect, and extend, with stable proportions, even spacing, and consistent stroke behavior in icon work.

Arrow 1.1 handles two operations. Text-to-SVG generation turns a prompt into one or more vector graphics. Vectorization converts an existing image, such as a sketch or fashion flat, into a clean, editable SVG. Both suit logos, iconography, illustrations, and technical drawings like engineering diagrams.

On AI Gateway, you call Arrow 1.1 with `generateImage` from the AI SDK. AI Gateway handles authentication, routing, retries, and failover, and this page lists live pricing for each operation.

## What to consider

Arrow 1.1 outputs SVG markup, not raster pixels. That's the point: you get scalable, editable files instead of fixed-resolution images. Plan your pipeline around vector output, and rasterize downstream if a workflow needs PNG or JPEG.

Pricing is per operation rather than per resolution tier. Generation and vectorization are billed separately; see : $0.2; : $0.15 for current rates. QuiverAI also notes that generating longer, highly detailed SVGs while keeping them clean and fully editable remains an open problem, so test dense compositions during evaluation.

## When to use

### Best For

- Logo and icon generation where output must stay scalable and editable
- Vectorizing sketches, fashion flats, and reference images into clean SVG files
- Illustrations and technical drawings that need precise alignment and consistent strokes
- Design pipelines that edit generated assets afterward, since output uses simple primitives

### Consider Alternatives When

- You need photorealistic raster imagery. `recraft-v4` targets natural skin, lighting, and textures
- You want raster generation with style presets plus optional vector output. `recraft-v3` covers both
- Your compositions are very long and dense. QuiverAI notes that highly detailed SVGs remain an open problem

## Best for

- **Logo and Icon Generation:** Scalable, editable vector output with stable proportions and consistent stroke behavior
- **Vectorizing Existing Images:** Converting sketches, fashion flats, and reference images into clean, editable SVG files
- **Technical Drawings:** Engineering diagrams and detailed line work that benefit from simple primitives and cleaner paths
- **Editable Design Assets:** Output built from shapes, text, and basic geometry that designers can extend

## Consider alternatives

- **Photorealistic Raster Imagery:** `recraft-v4` targets natural skin, lighting, and material textures for marketing and product work
- **Raster Plus Style Presets:** `recraft-v3` pairs 20+ style presets with native vector output
- **Long, Dense Compositions:** Highly detailed SVGs that stay clean and fully editable remain an open problem per QuiverAI

## Frequently asked questions

### What does Arrow 1.1 generate?

Arrow 1.1 generates SVG vector graphics, not raster pixels. Output is structured vector code built from shapes, text, and basic geometry, so files stay scalable and editable.

### What operations does Arrow 1.1 support?

Two: text-to-SVG generation and vectorization. Generation turns a prompt into vector graphics. Vectorization converts an existing image, such as a sketch, into a clean, editable SVG. Each operation has its own rate; see : $0.2; : $0.15.

### How does Arrow 1.1 differ from raster image models?

Raster models output fixed-resolution pixels. Arrow 1.1 writes vector code directly, making choices about control points, color, and gradients. You get resolution-independent files you can open and edit in any vector tool.

### What improved over the previous Arrow generation?

Design quality, prompt following, and SVG precision, especially in finer details. Arrow 1.1 also favors primitives over stacked paths, which makes output easier to edit, inspect, and extend.

### Is Arrow 1.1 a multimodal LLM?

No. Arrow 1.1 is a dedicated SVG generation model. Arrow 1.1 doesn't process conversation history or maintain a context window. Use `generateImage` from the AI SDK to send requests.

### How do I use Arrow 1.1 on AI Gateway?

Pass `quiverai/arrow-1.1` to `generateImage` from the AI SDK. AI Gateway routes the request, handles retries and failover, and returns the generated SVG.

### What does Arrow 1.1 cost?

Pricing is per operation, with separate rates for generation and vectorization. See : $0.2; : $0.15 and the pricing section on this page for current figures.

### Does Arrow 1.1 support zero data retention?

Zero Data Retention is not currently available for this model. Zero Data Retention is offered on a per-provider basis. See https://vercel.com/docs/ai-gateway/capabilities/zdr for details.

## Links

- [Model page](https://vercel.com/ai-gateway/models/quiverai/arrow-1.1)
- [AI Gateway documentation](https://vercel.com/docs/ai-gateway)
- [Provider model documentation](https://quiver.ai/blog/introducing-arrow-1-1/)


If you encounter incorrect information, can’t find what you need, need a capability Vercel doesn’t currently offer, or have any other feedback to report, please do it by following the instructions at https://vercel.com/feedback.md.
