Dreamina Launches a $50,000 Creative Contest Built Around Seedance 2.5's New Reference Tool
Dreamina launched a $50,000 creative contest built around Seedance 2.5's white model reference tool, giving every entrant the same 3D starting point. Here's how it works.
Dreamina launched "Story Unboxed," a global creative contest built entirely around one of Seedance 2.5's newer capabilities: white model reference generation. The premise is simple to state and genuinely interesting in practice, every entrant starts from the exact same 3D reference animation, and the contest is judging how differently creators can reinterpret it.
What a White Model Reference Actually Is
A white model is a simplified 3D scene, essentially an unfinished, untextured animation that retains only spatial structure, object volume, and motion relationships, the kind of rough previsualization block-out a VFX or animation team might build before committing to a final look. Seedance 2.5 can take that bare 3D reference and use it to drive character positioning, camera movement, and motion timing in a fully generated, styled video, letting a creator swap in entirely different characters, art styles, and settings while keeping the underlying blocking and motion identical.
That's a meaningfully different workflow than typing a text prompt and hoping the model interprets your intent. Reference-to-Video (R2V) generation using white models or green-screen footage gives creators precise control over spatial layout and character interaction, control that previously required either a full 3D pipeline or complex manual compositing to achieve.
How the Contest Actually Works
Dreamina provides an official white model reference video that every participant downloads and uses as their shared starting point. From there, entrants select Seedance 2.5 inside Dreamina, upload the reference, write their own creative prompt, and generate a stylized short film built on that identical spatial and motion foundation.
To qualify, entries must be published on at least one of TikTok, Instagram, or X with the required hashtags (#DreaminaAI, #DreaminaSeedance25, #DreaminaSeedance25Contest), and the submission form on Dreamina's site must be completed. All three steps are required for a valid entry.
Prizes and Timeline
Submissions run August 5–31, 2026, with review and judging in early September and winners announced mid-September. Ten jury-selected Grand Prize winners split a real cash pool: Champion takes $50,000 plus 1,000,000 Dreamina credits, runner-up gets $20,000 plus 100,000 credits, third place $10,000 plus 10,000 credits, and 4th through 10th place each receive a month of Dreamina Advanced Membership. A separate Social Media Likes Award gives the top 10 entries per platform (TikTok, Instagram, X) 30,000 Dreamina credits each, a real incentive independent of the jury results.
Why This Is a Smart Showcase, Not Just a Giveaway
Contests built around identical starting material are a genuinely effective way to demonstrate a specific tool's range, rather than a general model's overall quality. Since every entrant works from the same spatial blocking and motion, any creative variation in the final results comes entirely from prompting, styling, and narrative choices, not from different source material. That makes "Story Unboxed" function as a live, crowdsourced demo reel for white model reference control specifically, exactly the kind of feature Dreamina wants creators to understand and adopt.
Worth Knowing Before Entering
Seedance 2.5's broader rollout has been staged by account, region, and plan since its July 31 global launch, so access may vary even among people wanting to enter. The model itself supports 30-second clips at up to 4K, accepts up to 50 multimodal references, and includes localized region editing, real production-oriented capability well beyond the white model feature alone, though this specific contest is scoped narrowly around that one tool.
Competitive Context
Brand-run AI creative contests have become a fairly common adoption tactic across the AI video space this year, but most ask for open-ended prompts rather than a shared, identical starting reference. Dreamina's structure here is a more deliberate showcase of a specific, newer capability rather than a general promotional giveaway, which should produce a genuinely interesting side-by-side gallery of divergent creative interpretations once entries come in.
The Signal in the Noise
For filmmakers curious about reference-driven AI video generation specifically, this contest is a low-stakes way to actually test white model workflows against a real creative brief, with a meaningful prize pool as a bonus rather than the main draw. Worth remembering the entry requirements carry real friction, mandatory public posting with specific hashtags across name-brand platforms, which is standard for this kind of contest but worth factoring in before committing creative time to an entry.
