# Wan 2.7 Reference-to-Video

> Wan 2.7 Reference-to-Video 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-r2v
- Run it: https://artifio.ai/create/wan-2-7-r2v
- Modality: image to video
- Model brand: Wan
- Pricing: from about $0.080 per second (exact cost shown in the workspace before you confirm; varies with selected options)
- Typical generation time: ~300s (catalogue estimate)
- Full pricing table: https://artifio.ai/pricing/models

## About Wan 2.7 Reference-to-Video

Wan 2.7 Reference-to-Video (Wan2.7-r2v) is the reference-driven model in Alibaba's Wan2.7-Video release of April 2026. You supply reference material for a character, prop, or scene, and the model generates new shots that preserve that identity. It extends the reference-to-video mode Alibaba introduced in Wan 2.6, and the Model Studio documentation adds two concrete upgrades: multi-entity referencing and configurable voice timbre for the generated speech.

The wider Wan 2.7 generation adds cross-video consistency for up to five distinct characters, each with an exclusive voice tone and visual identity, which is the machinery this model exposes most directly. Artifio runs it in one workspace beside 100+ other models, billed per generation.

## Features

- Generates new video shots from reference material while preserving the referenced identity.
- Multi-entity referencing, so several characters or objects can be carried into one generated scene.
- Configurable voice timbre for generated dialogue.
- Outputs 720p or 1080p at 30 fps with audio.
- Any integer duration from 2 to 10 seconds.
- Part of a generation that holds up to five distinct characters consistent across separate videos.

## What people use it for

- Keeping a brand mascot or spokesperson identical across a whole campaign of generated clips.
- Placing a real person into scenes that were never filmed, voice included.
- Multi-character skits where two or more referenced identities share a frame.
- Product continuity shots where the referenced item must look the same in every scene.

## Reported strengths

- Identity preservation extends beyond faces to props and scenes, per Alibaba's documentation of multi-entity referencing.
- Voice timbre control means the character sounds consistent, and adjustable, across generations.
- No fine-tuning step; the reference material itself carries the identity.

## Reported limitations

- Duration is capped at 10 seconds, below the 15 seconds of the generation's text-to-video and image-to-video models.
- Resolution stops at 1080p.
- Hosted-only; no open weights exist for Wan 2.7 on the Wan GitHub organization.

## Sources

1. Alibaba unveils Wan2.7-Video to elevate creators from executors to directors (Alibaba Cloud): https://www.alibabacloud.com/blog/alibaba-unveils-wan2-7-video-to-elevate-creators-from-executors-to-directors_603009
2. Video generation, Alibaba Cloud Model Studio documentation (Alibaba Cloud): https://www.alibabacloud.com/help/en/model-studio/use-video-generation
3. Alibaba releases Wan2.7-Video AI model for multimodal video generation and editing (The Elec): https://www.thelec.net/news/articleView.html?idxno=6445
4. Wan2.2: open and advanced large-scale video generative models (GitHub (Wan-Video)): https://github.com/Wan-Video/Wan2.2

---

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-r2v for AI assistants and crawlers; the HTML page is canonical.