Lumenfall Pricing vs Pay Per Use: Two Ways to Pay Per Run

Spotlight search

Find any model, tool or page and jump to it.

All comparisons

Visual Sandbox vs Lumenfall

Both charge per run. One is an API. One is a place.

In short

  • Lumenfall charges the provider's public rate and adds nothing. No plan, no fee.
  • It is an API first. The Studio comes second, and the Playground needs a login.
  • We add a markup and give you a web app, a CLI, an MCP server and an API on one balance.
  • Both balances are prepaid and neither expires.

Our take. Use Lumenfall if you are a developer wiring models into your own product. Use Visual Sandbox if you want to make the thing.

See a model's price

Lumenfall routes 40+ image and video models through one OpenAI-compatible endpoint, at the providers' public rates with zero markup, and gives $1 free to start.

Pricing
The provider's public rate, no markup
Free plan
$1 of credit, no card
Also does
OpenAI-compatible SDK, failover, an Arena

Side by side

The same 9 questions, asked of Lumenfall

Visual Sandbox compared with Lumenfall, question by question
What is comparedVisual SandboxLumenfall logoLumenfall
What you pay forEach generation, at our listed priceEach generation, at the provider's public rate
What the balance is inMoney. One credit is one centMoney. A prepaid dollar balance
Free tierNone. A $10 minimum top up$1 of credit on signup, no card
Cheapest paid tierThere is no planThere is no plan. Top up any amount
A month you make nothing$0$0
The price before you runA dollar figure on every model page, before you sign inThe provider's price, on the model catalogue
Working as a teamNot offeredNot yet. Team tools are on the future premium list
From the terminal or an agentA CLI, an MCP server and an API, on the same balanceAn API. That is the product
The catalogue37 models across image, video, audio, 3D and vector40+ image and video models

Key takeaways

Is Visual Sandbox a good Lumenfall alternative?

Where we win

A place to make things
A web app with a library, shareable presets and every model page printing its price. Lumenfall is an endpoint.
Four doors, one balance
The web app, the CLI, the MCP server and the API spend the same cents.
Audio, 3D and vector too
Eleven Music at $0.19, Hunyuan 3D 3.1 at $0.63, Quiver Arrow at $0.25. Lumenfall lists image and video.
Skill packs for the agent
The MCP server hands an agent our prompting guidance before it prompts a model.
The balance does not expire
Neither does theirs. Both got this right.

Where Lumenfall wins

No markup
You pay what the provider charges. Their margin is a volume discount you never see.
One SDK for forty models
OpenAI-compatible. Send `size: 1024x1024` and it becomes whatever each model wants.
Automatic failover
When a provider goes down, the request routes around it.
A dollar to start
$1 of credit, no card, no minimum.
An Arena
Blind pairwise votes on model quality, with a public methodology.

Worth checking

Lumenfall is cheaper per run
By exactly our markup, on the models both carry. That is the honest headline.
It is for developers
Their founder says so on Reddit: 'our primary customers are developers, so the web UI is a bit technical.'
The Playground needs a login
lumenfall.ai/playground redirects to /app/playground. The price list is public; the try-it is not.
No team, no seat, no shared canvas
Listed as future premium features.
Image and video only
No audio, 3D or vector on the catalogue.

Two ways to pay per run

The price is on the page before the first run

This is the live price card off our Nano Banana 2 page. Lumenfall sells the same model at Google's public rate, through an API, and our price sits above it by our markup.

  • No account needed to read it, on either site.
  • Lumenfall's price is the provider's. Ours includes the web app, the CLI and the MCP server.
  • A quiet month costs $0 on both.

Lumenfall sells the call. We sell the place the call lands.

Nano Banana 2

What one run costs, before any account

1K
$0.09 / image
2K
$0.13 / image
4K
$0.19 / image

The money

Where the money goes

Both charge per run and neither bills a plan. The difference is what sits between you and the provider, and what you get for it.

$0.09

Visual Sandbox

Nano Banana 2 at 1K, including our markup and the app around it.

Provider rate

Lumenfall

The same model at Google's public price, through an API.

$1

Lumenfall, to start

Free credit, no card. Our minimum top up is $10.

RowVisual SandboxLumenfall
Price per runProvider rate plus our markupProvider rate
Plan or feeNoneNone
To start$10 minimum top up$1 free, then any amount
Web appYes, with a library and presetsA Studio, and a Playground behind login
CLI and MCPYes, on the same balanceNo
Audio, 3D, vectorYesNo
FailoverNoYes
TeamNoNot yet
  • Lumenfall's own page explains its money: volume discounts from providers, kept as margin, plus paid premium features later.
  • If you already have an OpenAI SDK client and your own front end, Lumenfall is the cheaper call. If you do not, the markup here buys the front end.
  • Their pricing page has a typo, 'GPR Image 1.5'. We mention it because a price list is only as good as its proofreading.

Which one fits

Which one fits you?

Four questions. The answer names Lumenfall when Lumenfall is the right call.

01 of 04

Are you building a product on top of the models?

Or use both

We cover the making. Lumenfall covers the shipping, if you ship software.

  1. 01

    Explore models and make the assets here, with the price on the page.

  2. 02

    Wire your own product to Lumenfall's endpoint at the provider's rate.

  3. 03

    Keep an agent on our MCP server for the audio, 3D and vector runs.

The call

Which one to pick

Pick Visual Sandbox when

  • You want a web app with a library and shareable presets.
  • You want a CLI and an MCP server on the same balance as the web app.
  • You want audio, 3D and vector next to image and video.
  • You want an agent handed prompting guidance before it prompts.
  • You want a per run cost you can put on an invoice, from a place a colleague can open.

Lumenfall logoPick Lumenfall when

  • You are a developer building a product with its own UI.
  • You want an OpenAI-compatible SDK over forty models.
  • You want automatic failover between providers.
  • You want the provider's rate with no markup.
  • You want to start with a dollar and no card.

Read the price, then run it

37 models across image, video, audio, 3D and vector. Each one prints its dollar price before you sign in.

See a model's price

Other comparisons

All comparisons

The provider's public rate per generation, with no markup, no platform fee and no monthly fee. You get $1 of credit on signup with no card, then top up any amount.