# Wan 2.7

> Wan 2.7 is available on Artifio.ai — a unified AI workspace with a single pay-as-you-go wallet.

- Canonical page: https://artifio.ai/models/wan-2-7-image
- Run it: https://artifio.ai/create/wan-2-7-image
- Modality: image editing
- Model brand: Wan
- Pricing: from about $0.025 per generation (exact cost shown in the workspace before you confirm; varies with selected options)
- Typical generation time: ~30s (catalogue estimate)
- Full pricing table: https://artifio.ai/pricing/models

## About Wan 2.7

Wan 2.7 (image) is the standard tier of Wan2.7-Image, the unified image generation and editing model Alibaba announced on April 1, 2026, a week ahead of the Wan2.7-Video suite. One model covers text-to-image, prompt-driven image editing, multi-image reference generation, and coherent image sets, which is why it appears here under editing: the same endpoint that generates can retouch, replace, and recompose existing pictures.

Alibaba's launch framing centered on precision: exact brand colors from color codes, print-quality text rendering, and character features that can be tuned down to bone structure and eye shape. On Artifio it runs beside 100+ other models in one workspace, billed per generation.

## Features

- Text-to-image, image editing, multi-image reference, and sequential image-set generation in one model.
- Output up to 2K (2048x2048) in every mode; 4K is reserved for the Pro tier.
- Accepts 0 to 9 input reference images; produces 1 to 4 images per run, or up to 12 in sequential set mode.
- Thinking mode is on by default for pure text-to-image, planning the composition before rendering.
- Interactive local editing with up to two bounding boxes per image to move, replace, or add elements.
- Color palette control from 3 to 10 specified colors, with codes and proportions honored in the prompt.
- Text rendering handles inputs up to 3,000 tokens across 12 languages, including formulas and tables.

## What people use it for

- E-commerce retouching, where a boxed region of a product photo is moved, replaced, or extended.
- Marketing images locked to exact brand color codes.
- Character sheets and storyboards generated as one coherent set of up to 12 images.
- Posters and academic figures with dense, correctly rendered text.

## Reported strengths

- Bounding-box editing gives spatial precision that prompt-only editors lack.
- Alibaba reports the model beat leading industry players on visual fidelity and text rendering in anonymized human preference tests.
- Character personalization goes to unusually fine detail, per the launch notes on bone structure and eye shape.
- One endpoint replaces separate generation and editing models.

## Reported limitations

- No 4K output on this tier; the documentation states 4K is not supported outside Pro.
- Thinking mode improves quality at the documented cost of longer generation times.
- Prompts are accepted in Chinese and English only, even though rendered text spans 12 languages.
- Interactive editing is limited to two bounding boxes per image.

## Sources

1. Alibaba unveils Wan2.7, redefining personalized and precision image creation (Alibaba Cloud): https://www.alibabacloud.com/blog/alibaba-unveils-wan2-7-redefining-personalized-and-precision-image-creation_602995
2. Wan image generation and editing 2.7 API reference (Alibaba Cloud): https://www.alibabacloud.com/help/en/model-studio/wan-image-generation-and-editing-api-reference

---

About Artifio.ai: 100+ AI image, video, and audio models in one workspace with a single shared wallet — no per-model subscriptions. This is the markdown rendition of https://artifio.ai/models/wan-2-7-image for AI assistants and crawlers; the HTML page is canonical.