Seedance 2.5 API

ByteDance introduced Seedance 2.5 at the Volcano Engine FORCE conference (reported June 23, 2026). The public API date has not been officially confirmed. This page is updated as availability changes; EvoLink will publish access once the official API documentation is public and the EvoLink route passes live validation.

Not yet through a public API. ByteDance typically ships new Seedance models to its own creative apps first, then opens API access via Volcano Engine and BytePlus. Until then, Seedance 2.0 is the current production model on EvoLink — and most 2.0 workflows should transfer to 2.5 with minimal changes (final parameters to be confirmed at launch).

As of July 2026, ByteDance has not published official per-second API pricing for Seedance 2.5 on Volcano Engine or BytePlus. The moment official rates are announced, this section lists them alongside the EvoLink route price — check back at launch.

EvoLink is pay-per-use with no subscription: you top up credits and each generation is billed per second of output video. Seedance 2.0 currently starts from about $0.07/s for Fast 480p video. See the full Seedance 2.0 pricing table →

Video API cost scales primarily with duration and resolution. Until official Seedance 2.5 rates land, a practical budgeting baseline is today’s Seedance 2.0 pricing: a 30-second clip costs roughly twice a 15-second clip at the same tier. Native 30-second generation also replaces multi-segment stitching, which cuts retries and total spend.

import requests, time

API_KEY = "YOUR_EVOLINK_API_KEY"
BASE_URL = "https://api.evolink.ai"
HEADERS = {"Authorization": f"Bearer {API_KEY}"}

# Works with Seedance 2.0 today — switch the model ID when Seedance 2.5 opens.
task = requests.post(
    f"{BASE_URL}/v1/videos/generations",
    headers=HEADERS,
    json={
        "model": "seedance-2.0-text-to-video",
        "prompt": "A 30-second brand story: sunrise over a coastal city, "
                  "drone push-in, upbeat soundtrack",
        "duration": 15,
        "quality": "720p",
        "aspect_ratio": "16:9",
        "generate_audio": True,
    },
).json()

while True:
    status = requests.get(
        f"{BASE_URL}/v1/tasks/{task['id']}", headers=HEADERS
    ).json()
    if status["status"] in ("completed", "failed"):
        break
    time.sleep(5)

print(status)

# 1) Submit the generation task (Seedance 2.0 pattern, Seedance 2.5-ready)
curl -X POST https://api.evolink.ai/v1/videos/generations \
  -H "Authorization: Bearer $EVOLINK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.0-text-to-video",
    "prompt": "A 30-second product story with synced audio",
    "duration": 15,
    "quality": "720p",
    "aspect_ratio": "16:9",
    "generate_audio": true
  }'

# 2) Poll the task until "status" is "completed"
curl https://api.evolink.ai/v1/tasks/TASK_ID \
  -H "Authorization: Bearer $EVOLINK_API_KEY"

The official Dreamina product page showcases Seedance 2.5 with up to 50 multimodal reference inputs across images, video, and audio (API-level limits pending confirmation) — beyond Seedance 2.0’s limits. Start organizing character sheets, product shots, motion clips, and audio cues now, and validate them on Seedance 2.0’s reference-to-video route today so your asset pipeline is ready early.

New EvoLink accounts receive trial credits that work across every model on the platform — including Seedance 2.5 when it launches. Sign up now, test your workflow on Seedance 2.0, and the same API key and credits will work with 2.5 once it launches on EvoLink. Sign up free and get credits →