Seedance 2.0 is now the default video generation model on CellCog - available to all users worldwide with no restrictions.
Two days ago we launched Seedance 2.0 with eligibility checks and regional gating. Today we’re deleting all of it: no more model choices, no eligibility checks, no geographic gating. Every video generation now runs on Seedance 2.0 automatically.
- Seedance 2.0 became the default video model on CellCog on April 9, 2026 - available to all users worldwide, individuals and businesses alike.
- The 1.5-to-2.0 selection step is gone: no model choices, no eligibility checks, no geographic gating.
- First-shot quality is the practical win: noticeably better motion, lighting, and fidelity means far fewer regenerations to reach your vision.
- Raw per-generation cost is 7-8x Seedance 1.5, but fewer iterations put the practical production cost around 3-5x - with a dramatically higher quality ceiling.
- Stricter content policies are handled automatically: violations retry up to 3 times, then fall back to Seedance 1.5 with a note.
- Lipsync mode is unchanged, and all five audio modes remain fully supported.
- What changed?
- Seedance 2.0 is now the default for every video generation on CellCog, worldwide - the selection process and all gating are gone.
- What does it cost?
- Raw generations cost 7-8x Seedance 1.5, but higher first-shot quality cuts iterations - practical production cost lands around 3-5x, with a much higher ceiling.
- Can I still use Seedance 1.5?
- Yes - ask for it explicitly when you want lower-cost generation, and the system falls back to it automatically on repeated content-policy blocks.
- Did lipsync change?
- No - lipsync continues on its dedicated engine, independent of the Seedance model, and all five audio modes remain supported.
§ 01What Changed
Cinematic quality by default. Seedance 2.0 delivers noticeably better motion, lighting, and visual fidelity. In our testing, it consistently produces great results on the first generation - fewer iterations to get to your vision.
Available globally. US, Japan, and all other regions are now fully supported. Individual consumers and businesses alike.
The cost reality, honestly. The raw per-generation cost is 7-8x higher than Seedance 1.5. However, 2.0’s higher first-shot quality means you typically need far fewer iterations to reach your final result. In practice, the apparent production cost is around 3-5x what it used to be - more expensive, but the quality ceiling is dramatically higher and better suited for production use. If you prefer lower-cost generation, you can still explicitly request Seedance 1.5.
Smart content policy handling. Seedance 2.0 has stricter content policies, especially around real people’s faces. When a violation occurs, the system automatically retries up to 3 times; if it still doesn’t pass, the agent falls back to Seedance 1.5 seamlessly and lets you know.
Lipsync unchanged. Lipsync mode continues on its dedicated engine (up to 30 seconds), independent of the Seedance model. All five audio modes remain fully supported: native AI audio, narration overlay, narration plus sound effects, lipsync, and silent generation.
§ 02Why Default to the Expensive Model
Because the honest unit of cost in video work is not the generation - it’s the accepted generation. A cheaper model that takes five takes to land a shot costs more than a premium model that lands it in one or two, and it costs something harder to price: your momentum. Making 2.0 the default optimizes for the number that actually matters.
§ 03The Honest Caveats
Work in progress, stated plainly: reference-to-video capabilities of Seedance 2.0 are not yet fully harnessed. We’re actively working on unlocking the full potential of the model, and you can expect further improvements in the coming updates. Content policy is also stricter than 1.5 - the automatic retry-and-fallback handles it, but face-heavy work will hit it more often.
Q1Why make the premium model the default?
Because iteration count is the real cost of video work. Seedance 2.0’s first-shot quality means you reach your vision in fewer generations - in our testing it consistently produces usable results on the first try, which changes the economics even at a higher per-generation price.
Q2What happened to the eligibility and region requirements?
Fully deprecated. The April 7 launch was gated to business accounts outside the US and Japan; as of April 9, every user in every region gets Seedance 2.0 automatically.
Q3How does the cost really compare?
Per generation, 7-8x Seedance 1.5. In practice, because you regenerate far less, the apparent production cost is around 3-5x - more expensive, but with a dramatically higher quality ceiling that’s better suited for production use.
Q4What happens when content policy blocks a generation?
The system retries automatically up to 3 times. If the block persists, the agent falls back to Seedance 1.5 seamlessly and tells you. Seedance 2.0 is notably stricter around real people’s faces.
Q5What about lipsync and audio modes?
Lipsync is unchanged - it runs on a dedicated engine supporting up to 30 seconds, independent of the Seedance model. All five audio modes remain supported: native AI audio, narration overlay, narration plus sound effects, lipsync, and silent.
Q6Is reference-to-video available?
At the time of this update it wasn’t yet fully harnessed - the model’s reference-driven capabilities were still being integrated into the video suite, with improvements shipping in subsequent updates.