CartVisualCreate account

Why Does FASHN Try-On Max Charge Differently When generation_mode Is Omitted?

For Try-On Max, omitting generation_mode does not mean selecting the lowest-cost fast mode. FASHN currently bills omitted mode as balanced at every resolution. Provider credits also depend on the selected resolution and num_images. Estimate the complete request before submission rather than assuming defaults are the cheapest configuration.

Published · By CartVisual editorial team

Identify the effective billing configuration

Omission means automatic mode selection. The current balanced billing rule does not prove which internal generation mode the service selects for a particular request.

A seller reviewing an API request may see only a product image, a model image, and a requested resolution. Because generation_mode is absent, the seller may assume no quality premium was selected. That assumption is not supported by Try-On Max's current pricing contract.

Inspect the exact request body. Record whether generation_mode was supplied, not merely what the application's interface displayed. If it was omitted, use the endpoint's documented automatic billing treatment. Then record resolution and num_images. These three values determine the initial provider-credit estimate.

The current FASHN Try-On Max reference states that omitted mode is billed as balanced across resolutions. Additional images multiply the cost. This is FASHN provider billing, not CartVisual credits or an estimate of the cost of a completed marketing campaign.

Calculate the request before authorizing it

Estimated provider credits = documented unit credits for the effective billing mode and resolution × num_images. This is a pre-submission estimate; observed usage must be checked separately.

Use the current pricing table in the official documentation, not a price copied from an older v1.6 integration. Select the row for the effective generation mode and the column for the requested resolution. Multiply the per-image figure by num_images.

The arithmetic is simple, but the request's origin can be ambiguous. A form may omit generation_mode while an internal configuration supplies a default. An integration may create several separate requests even though the seller sees one campaign task. The appropriate counting unit is the actual provider request, including its requested number of outputs.

Do not assume that every submitted request will produce an approved image. A credit estimate describes the configured operation; it does not predict quality, the number of revisions, or the final campaign's total expense.

Illustrative decision: a four-image jacket review

A fictional clothing merchant wants four candidate images of a jacket at 2K. The draft request specifies resolution='2k' and num_images=4, but leaves generation_mode absent.

Under the currently documented Try-On Max table, balanced at 2K is 3 provider credits per image. The illustrative request therefore has a listed estimate of 4 × 3 = 12 FASHN credits. If the merchant explicitly chose fast at 2K, the corresponding listed estimate would be 4 × 2 = 8 credits.

The difference is 4 provider credits for these hypothetical settings. It does not prove that fast would deliver an equally suitable image, and no request was submitted or billed.

Illustrative configurationListed unit creditsFour-image estimate
2K, omitted mode; balanced billing312
2K, explicit fast28
2K, explicit quality416

Decide whether the mode matches the task

Do not change the mode solely to make the estimate smaller. First define the work being purchased. Is the seller exploring whether the jacket source and model pose are suitable, or requesting candidates intended for detailed final review?

A fast request may be appropriate to consider for exploration. A higher-quality setting may be appropriate when the output itself is the intended review asset. Neither choice guarantees an acceptable garment depiction.

For v1.6-specific mode planning, use the existing performance-versus-quality guide. Its parameter names and recommendations must not be treated as the Try-On Max billing contract.

Close with an estimate-to-usage distinction

Record the endpoint, submitted mode or omission, effective billing interpretation, resolution, num_images, documentation date, and calculated estimate. Keep any later actual provider usage record separate from the pre-submission estimate.

If the application cannot show the final request settings, hold the cost estimate as incomplete. If the documented table changes, recalculate before another authorized request. Do not infer provider billing from CartVisual's separate product-creative workflow.

The task is complete when the seller knows which request configuration is being priced and why an omitted field does not necessarily mean the least expensive mode.

Sources

  • FASHN — Try-On Max API — Omitted-mode billing, resolution/mode credit table, and multiplicative num_images cost. Source checked 2026-11-16.

Scope and evidence

Prepared by CartVisual with AI assistance and checked against the listed sources. Illustrative estimates use the provider's documented November 16, 2026 table. No actual provider usage, billing, or image generation was observed.

Related decisions

Explore the AI virtual try on workflow