52 weekly newsletter header images: Ideogram 4.5 cost by tier on Sume

A year of weekly newsletter header images on Ideogram 4.5 costs $1.95 on the low tier, $3.90 on medium and $14.30 on high, at 52 images each. Those figures are the list prices ($0.03, $0.06 and $0.22 per image) times Sume's 1.25 margin, which gives $0.0375, $0.075 and $0.275 per image (read 2026-10-06). If you render three low-tier drafts and one high-tier final every week, the year comes to $20.15.

The number to plan around is not the per-image rate but your retake habit. A header is a single image with a headline in it, so the question is how many tries you need before the text is spelled right. The tables below let you pick the tier by that habit instead of by the cheapest sticker.

Drafts first, one high-tier final

The cheapest reliable pattern is to spend on the tier you would not want to retry. Three low-tier drafts cost $0.1125 per week. One high-tier final costs $0.275. Together that is $0.3875 a week, or $20.15 over 52 weeks, which is still less than rendering every header three times on high ($42.90).

Ideogram 4.5 takes quality as low, medium or high, and medium is the default when you omit it. The list price is the same for all sizes, so a 2K header costs what a 1K header costs on the same tier (read 2026-10-06). The resolution field is therefore a quality decision, not a price decision.

• Quote the headline in the prompt, in the exact words you want rendered.
• Pick the winner from the low-tier drafts, then re-send only that prompt on high.
• Keep the prompt text in your newsletter repo, so a retry next week starts from last week's winner.
• Read usage.cost on each image response: it is the USD amount that Sume billed to the wallet.

Compute the year in a few lines

The script below reproduces the table and the draft-plus-final plan. It uses the billed per-image rates, so you can change the weeks or the tiers and see the new total.

RATE = {"low": 0.0375, "medium": 0.075, "high": 0.275}
WEEKS = 52


def year(tier: str, tries: int = 1) -> float:
    return round(RATE[tier] * tries * WEEKS, 2)


def drafts_then_final(drafts: int = 3) -> float:
    weekly = RATE["low"] * drafts + RATE["high"]
    return round(weekly * WEEKS, 2)


for tier in RATE:
    print(tier, year(tier), year(tier, 3))
print("3 drafts + final:", drafts_then_final())

Which tier fits which header

Low is the right tier when the header is mostly colour, shape and a short title, because a misspelled word costs you $0.0375 and a re-roll. Medium is what you get when you omit quality, and at $0.075 it costs exactly what two low drafts cost. High is for the issue that goes on your site and into social cards, where small letterforms in the image will be seen at full size.

A practical rule is to keep a spend line for the newsletter in the same wallet as your video work. A single balance means a single place to read GET /v1/balance, and the usage ledger separates the image rows from the video rows, so a monthly review takes a minute instead of an afternoon.

What the ledger may round

Sume documents that its video and TTS prices round up to a whole cent for each job. For images the response gives usage.cost in dollars, and the Seedream example in the docs shows $0.04. Before you commit a yearly figure, submit one job on each tier and compare usage.cost with the rates above. If the ledger records whole cents for a single-image job, the three tiers become $0.04, $0.08 and $0.28, and the year becomes $2.08, $4.16 and $14.56.

Either way the ordering holds: one high-tier final is worth more than a second medium retry. Check the price table on the API pricing page for the current rate card before you budget.