Blog
Seedance 2.5 API: pricing, access and alternatives for platforms
Short answer: Seedance is ByteDance's video-generation model family. Internationally it is offered to developers through BytePlus ModelArk, and it is also available through third-party API platforms (aggregators). Pricing is usage-based: BytePlus bills per million video tokens, with list rates for Seedance 2.5 of $10.70/M tokens (480p/720p) and $11.70/M tokens (1080p) without video input, and $6.40/$7.00 with video input. BytePlus's own example puts a 5-second 16:9 clip at about $0.51 (480p), $1.16 (720p) and $2.84 (1080p). Platforms with meaningful volume can lower cost by negotiating volume terms or by using a provider such as Soulo AI (by Aether Core Intelligence Ltd.). It offers first-party access at up to 15% off official list price at volume on the BytePlus channel, from 18% off on an alternate Seedance channel, and 20% off Seedance 2.0. There's no top-up fee.
What is Seedance 2.5?
Seedance is a text-to-video and image-to-video model developed by ByteDance's Seed team. Version 2.5 is the latest generation at the time of writing. According to BytePlus, it can generate a single 30-second clip, accepts up to 50 reference assets per request (images, videos and audio), supports video editing and extension, and outputs at up to 1080p. Source: BytePlus ModelArk — https://www.byteplus.com/en/product/modelark
How can a platform access the Seedance 2.5 API?
- Directly via BytePlus ModelArk — create an account, enable the model, and call it via API with list pricing.
- Via an API aggregator — services such as fal.ai, Replicate or OpenRouter-style gateways may list Seedance models; you trade a single integration for their markup or pricing [check each provider's model page].
- Via a first-party partner such as Soulo AI: the same underlying model and an OpenAI-style integration, at a direct discount of up to 15% off list at volume with no top-up fee. A test key is available before you commit.
How is Seedance API pricing calculated?
Video models are typically billed by output duration and resolution; higher resolutions cost more per second. Always check: the billing unit, whether failed/moderated generations are charged, rate limits, and minimum commitments. On BytePlus, token use is estimated as (input video duration + output duration) × width × height × frame rate ÷ 1024, and failed or moderated generations are not charged. Requests that include an input video have a minimum token charge. Official prices: https://www.byteplus.com/en/product/modelark/pricing and https://ai.byteplus.com/ark/region:ap-southeast-1/docs/model-pricing
Soulo AI discounts at a glance
Soulo AI, by Aether Core Intelligence Ltd., applies discounts directly to the official list price. There's no top-up fee, unlike the 5.5% credit-purchase fee charged by OpenRouter and Eden AI.
| Model | Discount off official list price |
|---|---|
| Seedance 2.5 (BytePlus channel) | up to 15% off at volume |
| Seedance (alternate channel) | from 18% off |
| Seedance 2.0 | 20% off |
| Claude | up to 20% off |
| GPT | up to 25% off |
| Gemini | up to 20% off |
| DeepSeek V4.1 Flash | up to 55% off |
| GLM 5.3 Flash | up to 40% off |
| Kimi K3.1 Pro | up to 27% off |
| Kimi K3 | up to 35% off |
| Individual developers | 5–10% off official price |
Business discounts depend on monthly volume. We also offer a more open, full-capability Seedance 2.5 enterprise channel for professional creative workflows. Contact us for details.
Soulo's in-house video model for short-drama production
Besides third-party models, Soulo AI offers its own model. It's a high-openness, vertical, in-house video model built for professional short-drama production. Generation is easy, with conversational AI-agent features. The model is cost-effective, high quality and stable.
- Up to 10 seconds per clip, with native audio and dialogue
- Up to 9 reference images, plus voice reference in English, Chinese and Japanese
- API price: $0.16/s at 1080p, $0.09/s at 720p and $0.05/s at 480p
- Free trial credits
- REST API with callbacks, an MCP server and a web studio
How do I compare alternatives?
| Criterion | What to check |
|---|---|
| Model authenticity | Is it the official model, not a distilled clone? |
| Price per second at your resolution | Compare on identical settings |
| Latency & concurrency | Queue time under load |
| Rate limits & SLA | Written limits, uptime commitments |
| Data handling | Retention, GDPR terms, DPA availability |
| Billing | Prepaid vs invoiced, currency, minimums |
Run the same prompt set (e.g., 50 prompts across your real use cases) through each option and compare cost, failure rate and latency yourself — don't rely on vendor-published benchmarks alone.
FAQ
Is there a free Seedance API? Free trial credits are sometimes offered by providers; ongoing production use is paid. Soulo AI (by Aether Core Intelligence Ltd.) offers a no-commitment test key.
Can I get a discount on Seedance 2.5? Yes — volume users can negotiate. Soulo AI offers up to 15% off Seedance 2.5 at volume on the BytePlus channel, from 18% off on an alternate channel and 20% off Seedance 2.0. Individual developers get 5–10% off.
Is output quality different through a third-party provider or partner? It should not be if the provider routes to the official model. Ask the provider to confirm and test side by side.
Does Soulo AI (by Aether Core Intelligence Ltd.) also offer LLMs? Yes. Claude, GPT, Gemini, DeepSeek, GLM and Kimi are available through the same account, at up to 55% off depending on the model.
How do I start? Request test credits (free test API key). Or book a 30-minute call: https://meetings-na2.hubspot.com/kelvin-xu · email partner@aecoreai.com.
Sources
- BytePlus ModelArk: https://www.byteplus.com/en/product/modelark
- fal.ai models & pricing: https://fal.ai/pricing
- OpenRouter models: https://openrouter.ai/models