Arena ranking
| Model | Text-to-video Elo | Image-to-video Elo | Provenance |
|---|---|---|---|
| Wan 3.0 | 1,240 | Unranked | Artificial Analysis, August 2026 |
| Gemini Omni Flash | 1,237 | — | Artificial Analysis, August 2026 |
| Seedance 2.0 | 1,221 | 1,191 | Artificial Analysis, August 2026 |
| MiniMax H3 | 1,226 | — | Artificial Analysis, August 2026 |
Why Wan 3.0 has no image-to-video rank yet
The Artificial Analysis image-to-video arena ranks models on blind-vote preference for still-image animation. Wan 3.0 was added to the leaderboard within days of launch and is not yet ranked there. Seedance 2.0 leads at 1,191 Elo. See the image-to-video best-for hub for the full ranking.
Measured generation time
A 5-second 720p clip returns in [pending measurement] on the OneInfer measurement rig, averaged over [N] runs on [date] at [hardware]. The exact value is filled in once the run completes; the published methodology is preserved verbatim on this page.
Ready to test the workflow?
Create account & add creditsHow to read the table
- Elo is a relative score — only meaningful against other entries on the same leaderboard.
- Vendor-reported and independently measured results are shown as separate evidence classes.
- Harness version and clip duration matter; never quote a score without the harness name.
Frequently asked questions
Is Wan 3.0 the #1 text-to-video model?
Yes, as of August 2026. Wan 3.0 ranks first on the Artificial Analysis text-to-video arena at 1,240 Elo, ahead of Gemini Omni Flash at 1,237 and Seedance 2.0 at 1,221.
Why is Wan 3.0 not ranked on image-to-video?
Wan 3.0 was added to the Artificial Analysis image-to-video arena within days of launch and has not yet accumulated enough votes for a stable rank. Seedance 2.0 leads at 1,191 Elo in the meantime.
How fast is Wan 3.0?
Measured wall-clock latency is pending. The published methodology and the fill-in values appear on this page; refresh after the run completes.
Where is the canonical Wan 3.0 page on OneInfer?
The canonical model page with current pricing, providers, and capabilities is /models/alibaba/wan3.0-video. This page is a focused facet of that entity, not a replacement for it.
How should I treat benchmark or price claims?
Check each claim's provenance label and observed date. Vendor-reported and independently verified numbers are shown as separate evidence classes on this hub.
How do I call Wan 3.0 from the API?
POST to /v1/ula/generate-video with model "alibaba/wan3.0-video", a valid API key, and a prompt plus aspect_ratio and duration. Add first_frame, last_frame, or document_reference only when you need them.
Put Wan 3.0 to work
Fund a controlled evaluation, send a reference frame or document, and measure quality and cost on your own workload.