Seedance 2.0 Mini: ByteDance launches a video model twice as fast and half the price
Dreamina Seedance 2.0 Mini came out on June 17, 2026. Twice as fast as Seedance 2.0 Fast, at half the price. What it changes for creators who produce in volume.
On June 17, 2026, ByteDance released Seedance 2.0 Mini via Dreamina and CapCut. Not an overhaul, not a new foundation model. A lighter variant, positioned below Seedance 2.0 Fast, designed for those who generate a lot and fast.
Twice as fast as Fast. Half the price. The same inputs: text, image, reference video.
What Seedance 2.0 Mini does, concretely
The Seedance 2.0 range now has three tiers: Standard, Fast, Mini. Mini is the most accessible of the three.
The figures announced at launch:
- Speed: about 2 times faster than Seedance 2.0 Fast
- Image-to-video price: 0.023 RMB per 1,000 tokens (roughly 0.50 RMB per second in 720p)
- Accepted formats: text, image, video reference
- Resolution: up to 1080p depending on the mode
- Durations: 4 to 15 seconds per clip
The basic features stay intact. Mini is not a stripped-down version: it is a version optimized for speed and cost. The visual quality is presented as "comparable" to Fast, not inferior.
💡 Frank's Cut: On projects where I generate 30, 40 clips to find the 5 that work, speed matters more than perfection on the first try. Mini changes the cost/exploration ratio significantly.
How to access it
At launch, the most direct access is through Dreamina (the CapCut platform). The model appears as an option in the video generation interface. A dedicated API was listed as "coming soon" at the time of the June 17 announcement.
For users outside China, CapCut and Dreamina remain accessible via their web and mobile apps. The image-to-video input works well from an existing visual, with no complex configuration.
Why ByteDance is releasing Mini now
The logic is clear: Seedance 2.0 (released in February 2026) and Seedance 2.0 Fast were positioned on quality and versatility. Faced with competitors like Kling 3.0 Turbo (Kuaishou) and its fast, low-cost generations, ByteDance needed an option for creators who produce in volume.
Social content workflows, marketing teams, studios that generate dozens of variants per day have different requirements from directors looking for perfection on every shot. Mini targets exactly this segment.
It is also a direct answer to the positioning of Kling 3.0 Turbo, which had clearly targeted fast previewing and prototyping. The price war in AI video generation is only beginning.
What it changes for creative workflows
Here is what it really modifies:
Previewing and animatic. Generate 20 versions of a concept to choose the best camera movement, before investing credits on Standard or Fast. The cost cut in half makes this step economically viable.
Social content at volume. Reels, Shorts, TikToks. If you create 15 variants of the same product scene to test them, Mini changes the math.
Prompt testing. Before launching a long-duration or high-resolution generation, validate the prompt, the movement, the style on Mini first.
The limits to keep in mind. On complex scenes (crowds, precise physical interactions, specific textures), Fast or Standard will probably remain necessary. Mini is made for velocity, not for edge cases.
| Tier | Relative speed | Relative cost | Main use |
|---|---|---|---|
| Seedance 2.0 Standard | Base | Base | Maximum quality, complex scenes |
| Seedance 2.0 Fast | Faster | Cheaper | Daily production workflow |
| Seedance 2.0 Mini | 2x Fast | Half of Fast | Volume, proto, social |
Seedance in the AI video landscape of June 2026
The AI video market has restructured a lot in six months. Sora was discontinued from general public access on April 26, 2026. Kling 3.0 leads the leaderboards. Veo 3.1 holds the top of the basket on quality and native audio. And ByteDance is building a complete range to address every budget.
Seedance 2.0 Mini is one brick in this strategy, not a technical revolution. But it is the brick that was missing to make Seedance competitive on high-volume uses.
For creators already in the CapCut and Dreamina ecosystem, it is a direct and useful addition. For the others, it does not justify changing work environment, but it is worth noting when you compare the options.
The AI video generation tools listed here will be updated when the API is available.
FAQ
Foire aux questions
Réponses rapides aux questions les plus fréquentes sur cet article.
Is Seedance 2.0 Mini available outside China?
Yes. Dreamina and CapCut are accessible from countries where these apps are not blocked. Web access via dreamina.capcut.com works in most countries. The separate API was announced as "coming soon" at the time of launch.
What is the real quality difference between Mini and Fast?
ByteDance presents Mini as "comparable in quality to Fast" on the majority of use cases. In practice, creators report that on simple scenes and direct movements, the difference is small. On complex scenes with many dynamic elements, Fast or Standard are preferable.
Can Seedance 2.0 Mini replace Kling 3.0 Turbo?
Both are positioned on speed and cost. Kling 3.0 Turbo remains very strong on motion consistency and multi-shot. Seedance Mini has the advantage of being natively integrated into the CapCut/Dreamina ecosystem with editing tools. The best choice depends on your workflow.
Can you do image-to-video with Seedance 2.0 Mini?
Yes. Image-to-video is one of the supported modes. It is, in fact, one of the most common use cases for this tier: starting from an existing visual and animating it quickly to validate a concept.
Is there free access to Seedance 2.0 Mini?
CapCut generally offers a trial credit on new features. At the time of launch, the exact free-access conditions were not publicly specified for Mini. Check the Dreamina interface directly.
Does Seedance 2.0 Mini support native audio?
Seedance 2.0 supports synchronized audio generation since its 2.0 version. For Mini, the exact specifications on audio were not confirmed at launch. The market trend clearly pushes toward integrated audio, but it should be checked in the official documentation.