AI playing card deck art: 54 faces cost $1.62 to $11.88 on Ideogram

A 54-card deck (52 faces plus two jokers) is 54 image calls. At the Ideogram 4.5 list prices in Sume's docs, that is $1.62 at low, $3.24 at medium and $11.88 at high. Draw the rank and suit indices in code, so the art is the only thing you pay a model for.

Size does not change the Ideogram 4.5 price; quality does.

Cost of the deck by quality

The docs give the fal list price for ideogram/ideogram-v4.5 as $0.03, $0.06 or $0.22 per image by quality. Sume's endpoint pricing line is the amount charged to your wallet, so read cost_usd there and replace the numbers below with it.

Art in the model, indices in code

Ask the model for the card face's illustration inside a blank frame: no letters, no pips. Then draw the corner index (A, K, Q, J, 10 to 2) and the suit symbol with Pillow or HTML. That keeps every index the same typeface and avoids the case where a K comes back as an R on one card in fifty-four.

deck = 54
prices = {"low": 0.03, "medium": 0.06, "high": 0.22}
for quality, price in prices.items():
    print(f"{quality}: ${price * deck:.2f}")
print(f"drafts low + finals high: ${(prices['low'] + prices['high']) * deck:.2f}")

Running 54 calls

Submit with mode: "async" or a webhook rather than holding 54 blocking calls open: the images route waits up to 30 seconds and then returns a 202 job envelope, and slow configurations are the ones most likely to do that. Give each card its own Idempotency-Key (for example deck-v1-hearts-q), and keep status_url polling as the fallback. A queue_full response means your workspace's concurrency plus queue is full; wait for jobs to finish and retry.

What to check on delivery

Check card edges and the frame are consistent across the deck. If one face is off, regenerate that card only at high instead of re-running the deck; that is why the last row of the table is the one to plan around.