Video generation
Available now

Wan 2.7

by Alibaba

Alibaba's flagship video model.
Plans the shot before generating, with reference‑locked characters and accurate in‑frame text.

Wan 2.7

by Alibaba

Key features

Technical specifications

Up to 1080p

480p, 720p, 1080p tiers

27B MoE

Mixture-of-experts video DiT

T2V, I2V, Ref, Edit

Four core generation paths

Reasoning

Thinking-Mode prompt planning

Use cases

Cinematic short-form video

Direct multi-shot scenes with consistent characters, camera language, and lighting. Ad creative, brand films, and narrative shorts.

Character-consistent campaigns

Lock a subject from a reference image or clip, then generate multiple variations and shots without losing identity across the campaign.

Product and brand video

Use Wan 2.7's in-frame text and color control for product reveals, title cards, and ad creative that needs typography baked into the shot.

Iterative editing

Generate a base clip, then use instruction-based editing to refine wardrobe, lighting, or background without re-rolling the full generation.

Prompt examples

Cinematic shot

Slow dolly forward through a fog-bound forest at dawn, golden rays cutting between trees, a lone figure ahead, cinematic, 1080p

Edit prompt

Reference-led

Generate a video of this character walking through a rain-soaked Tokyo alley, neon reflections, mid shot, cinematic

Edit prompt

Product reveal

Studio-lit rotating reveal of a sleek minimalist watch on a marble slab, brand title appearing in clean typography, 5 seconds

Edit prompt

Overview

Wan 2.7 is Alibaba Tongyi Lab's flagship video generation model, the successor to the Wan 2.2 and 2.6 line. The headline shift is Thinking Mode, a planning step where the model reasons through prompt elements, composition, and motion before generating any frames. Combined with a 27B-parameter mixture-of-experts architecture and stronger reference-driven consistency, the result is cinematic 1080p video that holds together on complex multi-element briefs and across multi-shot sequences.

Simple pricing

Get started for free today, with the option to upgrade or cancel anytime.

Basic

$9/ month
billed as $0 per year

1100 monthly credits

1 user only

All models

Workflows

Standard

$24/ month
billed as $0 per year

3625 monthly credits

1 user only

All models

Workflows

Pro

$45/ month
billed as $0 per year

6350 shared monthly credits

1 user

+ up to 4 more at extra cost

All models

Workflows

Pro Max

$170/ month
billed as $0 per year

24650 shared monthly credits

1 user

+ up to 9 more at extra cost

All models

Workflows

Enterprise

For higher limits

Custom

pricing and billing terms

High-volume credits
Custom seat limits
All models
Workflows
Pricing Gradient

Free

For playing around

$0

forever free

Up to 20 credits
1 user only
Limited models
Workflows
Compare plan details

FAQs

What is Wan 2.7?
Wan 2.7 is Alibaba's flagship Tongyi Wanxiang video model. It introduces Thinking Mode, prompt reasoning before generation, on top of a 27B-parameter mixture-of-experts architecture, with cinematic 1080p output.
What is Thinking Mode?
Thinking Mode is a planning step before generation. Wan 2.7 reasons through prompt elements, composition, and motion before producing frames, which reduces artifacts and keeps complex briefs coherent.
How is Wan 2.7 different from Wan 2.6?
2.7 adds Thinking Mode, scales the architecture, improves reference-driven consistency, and renders in-frame text and color more reliably. Use 2.6 for fast generations, 2.7 when the prompt is complex or character consistency matters.
Does Wan 2.7 support reference-driven generation?
Yes. Pass in image or video references to lock subject identity across shots, useful for multi-shot campaigns and serialized content where the same character needs to appear consistently.
How do I use Wan 2.7 on Morphic?
Open Copilot, describe the shot or attach a reference, and pick Wan 2.7. For complex multi-element scenes, brand work with text in-frame, or character-consistent sequences, Wan 2.7 is the right pick over faster but less considered models.
How does Wan 2.7 compare with Veo 3.1, Seedance 2.0, and Kling 3.0?
Wan 2.7 leads on prompt reasoning and in-frame text/color control. Veo 3.1 leads on photorealism and 4K. Seedance 2.0 leads on native audio-visual generation and human motion. Kling 3.0 leads on multi-shot director-mode control. Pick by job.

Other models

Explore the rest of the Morphic model catalog.

See all models →

ChatGPT Images 2.5

OpenAI

OpenAI's image model, released September 2026. Sharper detail, precise edits, faster.

Lyria 3.5

Google DeepMind

Google's best-sounding music model. Full songs with structure, vocals, and lyrics.

MiniMax H3 Max Turbo

fal.ai

The fast tier of fal's post-trained H3. Twice the speed, nearly the same look.

Inworld TTS 2

Inworld

Ninety-five voices, 100+ languages. Speech that starts in under a second.