seedance-2.5-globalRead this first: Seedance 2.5 Global is suitable only for testing and for personal or small-studio use — it is not for production or commercial-scale workloads. It is an unstable channel. When the upstream is busy, a request can come back with an "overloaded" error instead of a video. Every failed job is refunded in full, so a failure costs you time, not money — but for anything that must succeed, use seedance-2.5. Seedance 2.5 Global is the fixed-shape, flat-price tier of Seedance 2.5: every request produces one 30-second clip at 720p and costs $2.00, whatever you send. The same 30-second 720p clip on seedance-2.5 bills at $0.270 per second — about $8.10 — so this is roughly a quarter of the price for the same length and resolution. That is the whole trade, and the things you give up are worth reading before you pick it. You do not choose the duration: 30 seconds is the only length, and a request asking for anything else is rejected at create time with a 400 rather than silently stretched to 30 and billed in full. You do not choose the resolution: 720p is the only output. There is no video editing and no video extension — sending a reference video returns a clear 400, because this tier simply has no such pipeline, and those are exactly the modes seedance-2.5 exists for. Real people are not supported. Input is text, or text plus up to 9 reference images given as public HTTPS URLs (base64 and data URIs are staged for you). Video and audio inputs are not accepted: a reference video or a reference audio track is rejected with a 400, never silently dropped. Six aspect ratios are accepted — 16:9, 9:16, 1:1, 3:4, 4:3 and 21:9 — which is actually one more than seedance-2.5 offers. Audio is always generated and there is no switch to turn it off; on seedance-2.5 you can set generate_audio to false, here you cannot, so a clip whose generated soundtrack trips the model’s own audio copyright check will fail and be refunded rather than come back silent. Plan on 6 to 11 minutes per clip — in our own testing one 30-second 720p generation finished in 369 seconds (a 12.5 MB H.264 file at 1280x720, 24 fps, stereo AAC), and a 9:16 one on 2026-10-01 took 652 seconds (720x1280, with an audio track). The limit that will bite hardest in production is not on the spec sheet: this tier is not built for high concurrency. Send requests one at a time or in small batches and let each finish; firing a burst at it is the fastest way to collect transient upstream failures. Pick Seedance 2.5 Global when you want a long 720p clip at a predictable flat cost and the fixed shape is fine. Pick seedance-2.5 when you need a specific length, 480p, editing, extension, real people, reference video or audio, or when you need to run many jobs at once.
Lock a character, product or style with up to 9 reference images — free of charge. Real people are not supported.
Generated video will appear here
Provide URLs and click Generate
An unstable channel, not for production or commercial-scale workloads. When the upstream is busy, a request can return an "overloaded" error instead of a video. Failed jobs are always refunded in full. For production work that must succeed, use seedance-2.5
Not a default you can override — it is the only output. Duration and resolution are not parameters here. A request asking for 10 seconds, or for 1080p, is rejected at create time with a 400 instead of being silently rewritten to 30s / 720p and billed the full $2.00
This tier generates from text and images only. There is no edit pipeline and no extend pipeline, so a reference video returns a clear 400 rather than being ignored. If you need to edit or lengthen an existing clip, that is seedance-2.5
Send jobs one at a time or in small batches and let each finish before queuing the next. Firing many requests at once is the fastest way to collect transient upstream failures — this tier has no capacity guarantee and is priced accordingly. For parallel or bursty workloads use seedance-2.5
There is no generate_audio switch on this tier. Every clip comes back with a stereo AAC track, and if the generated soundtrack trips the model’s own audio copyright check the job fails and is refunded — you cannot avoid it by asking for silence, only by changing the prompt away from music-heavy scenes. seedance-2.5 lets you set generate_audio to false
One price for every request, no per-second maths and no surprise on long jobs. The same 30-second 720p clip on seedance-2.5 bills at $0.270/s, about $8.10. Reference images are free
Images must be publicly reachable HTTPS URLs; base64 and data URIs are staged to public storage for you. Six aspect ratios — 16:9, 9:16, 1:1, 3:4, 4:3, 21:9 — one more than seedance-2.5 accepts. Real people are not supported. Reference video and reference audio are not accepted and return a 400
Async by design. In our testing one 30-second 720p generation finished in 369 seconds (12.5 MB H.264, 1280x720, 24 fps, stereo AAC) and a 9:16 one on 2026-10-01 took 652 seconds (720x1280, with audio). Poll every 10-15 seconds; only successful generations are charged
Seedance 2.5 Global is a Video Generation API provided by ByteDance. Read this first: Seedance 2.5 Global is suitable only for testing and for personal or small-studio use — it is not for production or commercial-scale workloads. It is an unstable channel. When the upstream is busy, a request can come back with an "overloaded" error instead of a video. Every failed job is refunded in full, so a failure costs you time, not money — but for anything that must succeed, use seedance-2.5. Seedance 2.5 Global is the fixed-shape, flat-price tier of Seedance 2.5: every request produces one 30-second clip at 720p and costs $2.00, whatever you send. The same 30-second 720p clip on seedance-2.5 bills at $0.270 per second — about $8.10 — so this is roughly a quarter of the price for the same length and resolution. That is the whole trade, and the things you give up are worth reading before you pick it. You do not choose the duration: 30 seconds is the only length, and a request asking for anything else is rejected at create time with a 400 rather than silently stretched to 30 and billed in full. You do not choose the resolution: 720p is the only output. There is no video editing and no video extension — sending a reference video returns a clear 400, because this tier simply has no such pipeline, and those are exactly the modes seedance-2.5 exists for. Real people are not supported. Input is text, or text plus up to 9 reference images given as public HTTPS URLs (base64 and data URIs are staged for you). Video and audio inputs are not accepted: a reference video or a reference audio track is rejected with a 400, never silently dropped. Six aspect ratios are accepted — 16:9, 9:16, 1:1, 3:4, 4:3 and 21:9 — which is actually one more than seedance-2.5 offers. Audio is always generated and there is no switch to turn it off; on seedance-2.5 you can set generate_audio to false, here you cannot, so a clip whose generated soundtrack trips the model’s own audio copyright check will fail and be refunded rather than come back silent. Plan on 6 to 11 minutes per clip — in our own testing one 30-second 720p generation finished in 369 seconds (a 12.5 MB H.264 file at 1280x720, 24 fps, stereo AAC), and a 9:16 one on 2026-10-01 took 652 seconds (720x1280, with an audio track). The limit that will bite hardest in production is not on the spec sheet: this tier is not built for high concurrency. Send requests one at a time or in small batches and let each finish; firing a burst at it is the fastest way to collect transient upstream failures. Pick Seedance 2.5 Global when you want a long 720p clip at a predictable flat cost and the fixed shape is fine. Pick seedance-2.5 when you need a specific length, 480p, editing, extension, real people, reference video or audio, or when you need to run many jobs at once. Through APIMODELS platform, you can access this model via a unified API with transparent pay-as-you-go pricing. Current pricing: 30s · 720p (the only option): $2, Reference images (up to 9): $0, For comparison: seedance-2.5 720p · 30s: $8.1.
Quickly generate brand promotion videos for ad campaigns and social media marketing.
Create compelling short-form video content for platforms like TikTok, Instagram, and YouTube.
Generate product feature demonstrations and tutorials to improve user conversion.
Produce course explanations, knowledge explainers, and training videos at an affordable price.
Seedance 2.5 Global is available through APIMODELS at: 30s · 720p (the only option): $2, Reference images (up to 9): $0, For comparison: seedance-2.5 720p · 30s: $8.1. Billing is pay-as-you-go — you only pay for what you generate.
Sign up at APIMODELS, get your API key, and call our unified API endpoint. We provide detailed API documentation with code examples in cURL, Python, and Node.js.
APIMODELS offers the same Seedance 2.5 Global model through our aggregation platform. We provide a unified API interface so you do not need separate accounts for each provider - one API key to access all models.
On APIMODELS, Seedance 2.5 Global runs alongside 60+ models on one API key and one balance, so choosing is about fit, not lock-in. It supports ⚠️ Testing & Personal Studio Use Only, ⚠️ Unstable — May Fail, Refunded, 30s Fixed — No Other Length, 720p Fixed — No Other Resolution, $2.00 Flat per Video, Text or up to 9 Ref Images, 6 Aspect Ratios, ⚠️ No Video / Audio Input, ⚠️ No Edit / No Extend, ⚠️ No Real People, ⚠️ Audio Always On, ⚠️ Not for High Concurrency, ~6–11 min / Clip, and you can weigh it on price and capability against other Video Generation models, then switch by changing a single model-name string — no new account or integration. Browse every Video Generation option with live pricing at apimodels.app/models.
Seedance 2.5 Global supports: ⚠️ Testing & Personal Studio Use Only, ⚠️ Unstable — May Fail, Refunded, 30s Fixed — No Other Length, 720p Fixed — No Other Resolution, $2.00 Flat per Video, Text or up to 9 Ref Images, 6 Aspect Ratios, ⚠️ No Video / Audio Input, ⚠️ No Edit / No Extend, ⚠️ No Real People, ⚠️ Audio Always On, ⚠️ Not for High Concurrency, ~6–11 min / Clip. See the APIMODELS docs for full parameters and call examples.
Yes. APIMODELS exposes Seedance 2.5 Global through a single unified API and one key — no separate provider accounts, and no need to handle each provider's regional network access yourself.
We support Stripe (Visa, Mastercard, and other international cards) and Alipay. Credits are available instantly after payment.