Seedream 5.0 Lite 9MP (3072x3072) vs Sume's resolution tiers
fal lists Seedream 5.0 Lite up to 9 megapixels at $0.035. Sume takes tiers (512, 1K, 2K, 4K), so read the model's resolution descriptor before pinning one.

fal lists Seedream 5.0 Lite at $0.035 per image with output up to 9MP (3072x3072). Sume's API does not take a megapixel number: resolution is a tier (512, 1K, 2K or 4K), and a model accepts only the tiers its catalog descriptor lists. So bytedance-seed/seedream-5-lite may or may not offer a tier that reaches 3072 on a side; read supported_parameters.resolution for it before pinning one.
The model facts are from fal's Seedream 5.0 Lite page, read 2026-10-03. The request rules are from the Image API docs.
What fal's page says
Seedream 5.0 Lite is described for product mockups, marketing assets and creative iteration. It takes detailed natural-language prompts, returns PNG via URL or data URI, and ships with a content filter on by default that the platform lets you switch off in controlled setups.
| Item | Value on fal's page |
|---|---|
| Price | $0.035 per image |
| Maximum output | 9MP (3072x3072) |
| Images per generation | 1 to 6 |
| Output | PNG via URL or data URI |
Pinning a tier on Sume
The docs say Sume normalises resolution to 512, 1K, 2K or 4K, and that a request using a value the model does not list is rejected with 400 unsupported_parameter rather than silently dropped. The script below prints the resolution tiers each catalog row declares.
import os
import requests
key = os.environ["SUME_API_KEY"]
url = "https://api.sume.com/v1/images/models"
resp = requests.get(url, headers={"Authorization": f"Bearer {key}"}, timeout=30)
resp.raise_for_status()
for model in resp.json()["data"]:
tiers = model["supported_parameters"].get("resolution", {}).get("values", [])
print(model["id"], tiers)
Limits
Do not assume the fal 9MP ceiling carries over: Sume's own docs say its catalog rows are what governs a request, and fal's content-filter switch is not exposed by Sume's schema. Treat the catalog as the contract.
Sources
Related posts
More in Models
- First frame plus style sheet in one video call: frame_images wins
On Sume /v1/videos, frame_images plus input_references runs as image-to-video, so plan on your style sheet being set aside. Split the call and pin the opening.
- Shortest AI video clip via API: 2 seconds on Wan 3.0 only
Sume's video catalog by minimum duration: Wan 3.0 takes 2 seconds, Omni 3, Seedance and Kling 4, MiniMax H3 5. Plus how to trim below a floor.
- Shortest AI video clip by model: 2, 3, 4 or 5 second floors on Sume
Wan 3.0 starts at 2 seconds; Gemini Omni Flash at 3; Seedance and Kling at 4; MiniMax H3 at 5. The minimum duration for each Sume video model.
- Shortest AI video clip each Sume model accepts, and its price
Wan 3.0 starts at 2 seconds ($0.125 at 480p), Omni Flash at 3, Seedance 2.5 and Genjutsu at 4, H3 at 5. Minimum request per model, priced.
Written by Sume