ByteDance’s Seedance 2.0 Just Took the Image-to-Video Crown
The image-to-video leaderboard has a new king. dreamina-seedance-2.0-720p from ByteDance’s Dreamina lab has overtaken gemini-omni-flash to claim the #1 spot — and a new open-weight contender from Alibaba just cracked the top 10.
What Changed Today (June 24, 2026)
- 🏆 New #1 in Image-to-Video: dreamina-seedance-2.0-720p overtook gemini-omni-flash for the top position on the Arena leaderboard.
- 🆕 New model in top 10: wan2.7-i2v (Alibaba) entered the image-to-video top 10 with an ELO of 1434.
Image-to-Video: A Major Shakeup at the Top
dreamina-seedance-2.0-720p from ByteDance’s Dreamina team is now the top-rated image-to-video model on the Arena leaderboard, dethroning gemini-omni-flash. This is a significant shift — it marks the first time a ByteDance model has held the #1 position in this category, breaking a stretch of dominance that rotated between xAI’s Grok Imagine Video and Google’s Veo family. Seedance 2.0 is the successor to ByteDance’s earlier Seedance line, and its Arena performance suggests it handles motion coherence, prompt fidelity, and visual quality at a level users consistently prefer over the competition.
For developers looking to integrate, watch for availability via ByteDance’s API or third-party inference providers — official API IDs and pricing have not yet been widely published, so early adopters should check Dreamina’s developer docs directly. If you’ve been building on Grok Imagine Video or Veo 3.1 for image-to-video pipelines, this is worth benchmarking against your specific use cases.
FREE GUIDE
Stop Writing Design Specs by Hand
Get the free visual guide: how AI tools generate GAMP 5 documentation directly from your PLC and DCS exports. Used by Life Sciences engineers who are done doing it manually.
No spam. Unsubscribe anytime.
Meanwhile, wan2.7-i2v from Alibaba quietly entered the top 10 with an ELO of 1434 — which is notably higher than several models that were previously entrenched in the rankings, including the prior leader from the reference data. This is the next generation of Alibaba’s Wan series (succeeding wan2.5-i2v), and as an open-weight model, it’s a strong option for teams that need self-hosted or customizable image-to-video inference. Expect it to be available via Alibaba Cloud’s Model Studio or through community deployments on Hugging Face. Use the API model ID wan2.7-i2v where supported.
Current Leaders at a Glance
| Category | #1 Model | Provider | Score |
|---|---|---|---|
| Image-to-Video | dreamina-seedance-2.0-720p | ByteDance (Dreamina) | New #1 (ELO TBC) |
So What?
If you’re building image-to-video features into your product, today’s changes demand attention. ByteDance’s Seedance 2.0 taking #1 means the competitive field has expanded beyond the Google-xAI duopoly that dominated for months — and with Alibaba’s wan2.7-i2v entering the top 10 at a very competitive ELO of 1434, open-weight options are now genuinely rivaling proprietary APIs on quality. The practical move: run your own evals against Seedance 2.0 and Wan 2.7 this week. If Seedance 2.0’s API access and pricing are favorable, it could be your new default. And if you need to self-host for cost, latency, or data privacy reasons, Wan 2.7 just became the strongest open model in the category. The image-to-video space is moving fast — lock in your benchmarks now before the next shakeup.
