Private AI
Unified Wan 2.7 video model. One model id that supports text-to-video, image-to-video, reference-to-video, or video editing based on what you upload.
Added Apr 3, 2026
Approx. Price
$0.200 per video
Model Type
both
Generation controls available for this model.
Output Format
Default Duration
5
14 duration options
Aspect Ratio
Default
16:9
Options (5)
16:9 (Landscape), 9:16 (Portrait), 1:1 (Square), 4:3 +1 more
Used for text-only generations and optional for reference/edit flows.
Duration
Default
5
Options (14)
2 seconds, 3 seconds, 4 seconds, 5 seconds +10 more
2-15 seconds. Reference-to-video and video edit cap at 10 seconds.
Mode
Default
auto
Options (3)
Auto-detect, Edit Video, Reference to Video
Auto-detect uses your uploads. Edit Video requires a video. Reference to Video requires a reference video, supports one image total, and does not support custom audio.
Negative Prompt
Default
N/A
Describe visual details to avoid in the generated video.
Human preference benchmarks sourced from LMArena.
Text to Video
#11 / 42
Arena Score
1347.9
Votes
8,092
Confidence Interval
1337.9 - 1357.8
Published 2026-07-07 · Matched as wan2.7-t2v
LMArena DatasetPrompt Expansion
Default
true
Options (2)
true, false
Use AI to expand your prompt
Resolution
Default
720p
Options (2)
720p, 1080p
Output resolution