Model + Tool

Wan 2.7 Text-to-Video Generator

What it is

A Wan text-to-video generator uses Alibaba's open-source Wan 2.7 (or 2.6 / 2.5) model to produce short video clips — the most capable open-weights video model available, with strong character animation and native Chinese + English support.

Try it now

Or open the full tool →

Output specs

Resolution

720p / 1080p

Duration

5–8 seconds

Aspect ratios

16:9, 9:16

Cost

From ~12 credits per 5s clip

Features

✓Wan 2.7, 2.6, 2.5
Three Wan generations on Mobbi — 2.7 latest, 2.6 stable, 2.5 budget. Switch with one click.
✓Open-source weights
Wan is the most capable open-weights video model in 2026. Mobbi runs it on managed infra so you skip GPU setup.
✓Strong character animation
Wan 2.7 trains on Chinese drama content; character motion and facial expression are particularly strong.
✓Bilingual prompting
Native Chinese + English support. Mix tokens for stylistic control.

Best for

→Workflows that prefer open-weights models for licensing reasons
→Character-driven scenes (Wan 2.7 trained heavily on this)
→Cost-conscious production where Wan beats Sora 2 on price
→Cross-market campaigns mixing English and Chinese prompts

Pro tips

1.Wan 2.7 is the recommended default; 2.6 is the stable fallback. Use 2.5 only for cost-sensitive batch work.

2.Audio import mode (Wan 2.7) lets you pass audio for lip sync. Worth using vs separate lip-sync step.

3.Wan rewards detailed Chinese cinematography terms ("电影感", "国漫风") if you want a specific aesthetic.

FAQ

Is Wan open source?+

Yes — Alibaba released Wan model weights publicly. Mobbi runs Wan via managed infrastructure so you don't need GPUs.

Related on Mobbi

Sora 2 Text-to-Video →Kling Text-to-Video →Hailuo Text-to-Video →

Try other model + tool combos

Sora 2 Text-to-Video Generator Kling 2.6 Text-to-Video Generator Hailuo 2.3 Text-to-Video Generator Sora 2 Image-to-Video Generator Veo 3 Text-to-Video Generator Veo 3 Image-to-Video Generator

Browse all generators →