The cheapest text-to-image APIs in 2026

FLUX 1.1 [pro] has the lowest listed per-image price on Infer, at $0.02/image, though its own page also shows a conflicting $0.04/img mention that's never been reconciled. Seedream 4.0 is the cheapest price on the list backed by a confirmed leaderboard rank, at $0.03/image. Below both, Nano Banana 2 ($0.039), GPT Image 1.5 ($0.04), Imagen 4 Standard ($0.05), and Ideogram 3.0 ($0.06) fill out Infer's text-to-image ladder, all observed July 22, 2026. Every one of these bills only successful generations: a failed render costs nothing.

Start generating on Infer — pay per use, no minimums →

Ranked by price: text-to-image APIs across providers

*FLUX 1.1 [pro]'s primary pricing block reads $0.02/image, but a separate mention on the same page cites $0.04/img. Neither figure has been reconciled as of this writing; confirm the live rate in the console before quoting it to a client.

Infer hosts all six of these models and takes the same cut regardless of which one wins this ranking. We have no favorite here; the table above is just the rate card.

What the ranking actually buys you

Seedream 4.0's $0.03/image is the number to build a budget around if benchmark standing matters as much as the bill: it's the cheapest model on this list with a confirmed Infer image-leaderboard position, #6 at 1198 Elo. FLUX 1.1 [pro] is nominally cheaper, but its price is unresolved and it carries no published Elo on Infer's leaderboard at all. That's a real gap, not a rounding error, when a client asks "why this one."

The off-Infer column tells a second story. Seedream 4.0 is the one model here where the outside price matches Infer's almost exactly ($0.03 on fal.ai too), which is the strongest confirmation on this table. Nano Banana 2 and Imagen 4 Standard don't line up as cleanly: Google's Gemini API reportedly bills Nano Banana 2 by output resolution rather than a flat per-image rate (per aggregator sources, unverified against Google's own pricing page), so Infer's $0.039 flat price and Google's $0.045-$0.151 tiered price aren't quoting the same product shape, only a rough neighborhood. Imagen 4 Standard's gap is sharper: Infer's $0.05 sits squarely between an aggregator-reported Google Standard ($0.04) and Ultra ($0.06) tier, neither confirmed against Google's own pricing page, and that discrepancy hasn't been explained on either page.

What a real image job costs

Same six models, three job sizes, at Infer's listed per-image rates.

*FLUX 1.1 [pro]'s range reflects the unresolved $0.02 vs. $0.04 pricing discrepancy on its own Infer page, not a resolution or volume tier. At the low end it's the cheapest option on every row; at the high end it lands roughly level with GPT Image 1.5.

The spread between cheapest and priciest confirmed models never exceeds 2x on this list (Seedream at $0.03 against Ideogram at $0.06). Whichever figure turns out to be true for FLUX 1.1 [pro], it changes which end of that spread it sits at, not whether it's competitive.

What affects the price

Resolution is the biggest lever, but only some of these models charge for it. Google's own Gemini API prices Nano Banana 2 in four resolution tiers, from $0.045 at 512px up to $0.151 at 4096px, plus a 50% batch-API discount at every tier, per aggregator sources. Infer's $0.039 rate for the same model doesn't split by resolution; it's a single flat price. Imagen 4 works the same way on Google's side, three tiers (Fast/Standard/Ultra) instead of resolution steps per aggregator sources, while Infer again lists one flat number.

Volume matters on exactly one model here: FLUX 1.1 [pro]'s Infer page references a discount past 10,000 requests/month. None of the other five document an equivalent tier, so don't assume one exists for Seedream, Nano Banana, GPT Image, Imagen, or Ideogram just because FLUX has it.

Billing mechanics apply uniformly. Every model in this table runs on Infer's pay-per-successful-generation model: a request that errors or times out isn't billed, which matters more at the 10,000-image end of the table above than at the 100-image end.

FLUX or Seedream: which cheap price to trust

Choose Seedream 4.0 if the number on the page has to hold up when someone checks it: $0.03/image, confirmed on Infer's own pricing block and matched almost exactly by fal.ai's independent listing. Choose FLUX 1.1 [pro] only after confirming the live rate in the console, since its own page can't agree with itself on whether the price is $0.02 or $0.04. If the job needs in-image text to actually read correctly, neither is the pick — Ideogram 3.0 costs twice Seedream's rate but is the one model here built specifically for that job, and Nano Banana 2 at $0.039 is the cheaper typography-capable alternative if Ideogram's price is the sticking point.

Related reading

• The best text-to-image models in 2026, ranked: the same six models, ranked on quality and use case rather than price.
• GPT Image 1.5 vs Nano Banana 2: a head-to-head on the two mid-priced models above.
• The cheapest AI video generation APIs in 2026: the same roundup shape, applied to video.
• The AI video pricing index: the master price table across every model and provider on Infer.
• All pricing and cost breakdowns: every calculator and cost page in one place.