ByteDance

Seedance 1.0

ByteDance's video generation family, in Pro and Lite variants, covering text-to-video and image-to-video at 480p and 720p. Available as a hosted endpoint; weights are not publicly released.

Developer ByteDance
Task text-to-video · image-to-video
Resolutions 480p · 720p
Weights not public

Capabilities

Characteristics reported by ByteDance in the Seedance 1.0 model introduction. Treat them as vendor claims, not independent measurements.

Multi-shot coherence

Subject appearance and scene layout are held stable across shot changes within a single generation, rather than each cut resampling the subject.

Motion and physics

The Pro variant is tuned for plausible trajectory and contact behaviour — the difference most visible in sports, vehicles and cloth.

Prompt adherence

Handles compound prompts that specify subject, action, camera move and style together, instead of collapsing to the dominant noun.

Style range

Covers photographic output through to stylised and animated looks from the same checkpoint.

Reported evaluations

ByteDance publishes results on its internal SeedVideoBench-1.0 suite and points to the public Artificial Analysis arena. Both charts below are ByteDance's own; we have not reproduced the numbers, and the arena leaderboard changes over time, so check the live ranking before quoting it.

SeedVideoBench-1.0 multi-dimensional evaluation — **SeedVideoBench-1.0.** ByteDance's internal evaluation across instruction following, motion and visual aesthetics.

Artificial Analysis arena results — **Artificial Analysis.** Blind pairwise preference scores from the public arena, as reported at release.

Variants

Pro targets output quality; Lite trades some fidelity for a lower cost per run. Per-run pricing is listed on each model page — it changes, so we don't mirror it here.

Endpoint	Variant	Task	Resolution
`bytedance/seedance-v1-pro-t2v-480p`	Pro	Text-to-video	480p
`bytedance/seedance-v1-pro-i2v-480p`	Pro	Image-to-video	480p
`bytedance/seedance-v1-pro-i2v-720p`	Pro	Image-to-video	720p
`bytedance/seedance-v1-lite-t2v-480p`	Lite	Text-to-video	480p
`bytedance/seedance-v1-lite-i2v-480p`	Lite	Image-to-video	480p
`bytedance/seedance-v1-lite-i2v-720p`	Lite	Image-to-video	720p

Run it

Swap the endpoint path for any variant in the table above; the request body is otherwise the same.

# 1. submit the job
curl -X POST "https://api.wavespeed.ai/api/v3/bytedance/seedance-v1-pro-i2v-480p" \
  -H "Authorization: Bearer $WAVESPEED_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "image": "https://example.com/still.jpg",
        "prompt": "The camera pushes in slowly as steam rises from the cup",
        "duration": 5,
        "enable_sync_mode": false
    }'

# -> {"code": 200, "data": {"id": "<request-id>", "status": "created", ...}}

# 2. poll until status is "completed"
curl "https://api.wavespeed.ai/api/v3/predictions/<request-id>/result" \
  -H "Authorization: Bearer $WAVESPEED_API_KEY"

# -> {"code": 200, "data": {"status": "completed", "outputs": ["https://..."]}}

import os, time, requests

API = "https://api.wavespeed.ai/api/v3"
KEY = os.environ["WAVESPEED_API_KEY"]
HEADERS = {"Authorization": f"Bearer {KEY}"}

# submit
res = requests.post(
    f"{API}/bytedance/seedance-v1-pro-i2v-480p",
    headers={**HEADERS, "Content-Type": "application/json"},
    json={
      "image": "https://example.com/still.jpg",
      "prompt": "The camera pushes in slowly as steam rises from the cup",
      "duration": 5,
      "enable_sync_mode": false
    },
    timeout=30,
)
res.raise_for_status()
request_id = res.json()["data"]["id"]

# poll
while True:
    data = requests.get(
        f"{API}/predictions/{request_id}/result",
        headers=HEADERS,
        timeout=30,
    ).json()["data"]

    if data["status"] == "completed":
        print(data["outputs"][0])
        break
    if data["status"] == "failed":
        raise RuntimeError(data.get("error", "generation failed"))
    time.sleep(1.5)

const API = "https://api.wavespeed.ai/api/v3";
const KEY = process.env.WAVESPEED_API_KEY;
const headers = { Authorization: `Bearer ${KEY}` };

// submit
const submit = await fetch(`${API}/bytedance/seedance-v1-pro-i2v-480p`, {
  method: "POST",
  headers: { ...headers, "Content-Type": "application/json" },
  body: JSON.stringify({
      "image": "https://example.com/still.jpg",
      "prompt": "The camera pushes in slowly as steam rises from the cup",
      "duration": 5,
      "enable_sync_mode": false
    }),
});
const { data: { id } } = await submit.json();

// poll
for (;;) {
  const res = await fetch(`${API}/predictions/${id}/result`, { headers });
  const { data } = await res.json();

  if (data.status === "completed") {
    console.log(data.outputs[0]);
    break;
  }
  if (data.status === "failed") throw new Error(data.error ?? "generation failed");
  await new Promise((r) => setTimeout(r, 1500));
}

Requests are asynchronous: POST returns a request id, then you poll /predictions/<id>/result until status is completed. Set enable_sync_mode: true to have the call block and return outputs directly.

API keys are created in the WaveSpeed dashboard.

Open Seedance on WaveSpeed API reference

Limitations

Worth knowing before you build on it: output is synthetic and unsuitable as a source of factual imagery; results inherit biases present in the training data; image-to-video quality tracks input image quality closely, and small changes to a prompt or seed can move the result noticeably. Each endpoint is fixed to the resolution listed above.