flux-2-pro vs flux-2-flex on Sume: price, limits and a 100-image bill

flux-2-flex costs $0.0625 per image on Sume and flux-2-pro costs $0.0375, so Flex is 1.67x Pro. The catalog gives them the same shape: up to 10 input_references, n up to 4, and png, jpeg or webp output.

Because the interface is identical, you can switch by changing one string and measure the difference on your own prompts rather than guess.

Bill comparison

A/B in one script

Run the same prompts on both and keep the output URLs side by side. Two images per call keep the request count low.

import os, requests

H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
prompt = "Isometric desk setup with a mechanical keyboard, pastel palette"
for model in ["black-forest-labs/flux.2-pro", "black-forest-labs/flux.2-flex"]:
    r = requests.post("https://api.sume.com/v1/images", headers=H, timeout=60,
                      json={"model": model, "prompt": prompt, "n": 2})
    print(model, r.status_code)
    if r.status_code == 200:
        print(r.json()["usage"]["cost"], [d["url"] for d in r.json()["data"]])

How to decide

Take a fixed set of 20 prompts, run both rows, and count the images you would actually ship. If Flex ships noticeably more usable images per dollar, take it; if not, the cheaper row wins. Sume does not publish a quality ranking, so this test is yours to run.

For lettering inside images see which Sume image model for text in images.

Cost of an experiment

A 20-prompt, 2-model, 4-image test is 160 images: $8.00 in total. That is the price of knowing which row to use for the next 10,000.

• Fix the prompts and the aspect ratio so only the model varies.
• Reruns of the same prompt can differ, so judge a row on several prompts, not one.
• Read GET /v1/images/models/{model_id}/endpoints for the live pricing line before a large run: prices are catalog data and can change.