Pay Only for What You Use
Prepaid balance in US dollars, no subscription. Every model shows its price before you run it, here and on its model page.
As a small business under Section 19 of the German VAT Act (UStG), no VAT is charged or shown.
- 1 credit
- $0.01Prepaid, billed in USD
- Top up
- $5–$500Any whole amount, by card
- Models with a price
- 274Price shown before every run
- Google sign-up
- 10 creditsTrial, usable after 24 h
Estimate
Estimate your cost
Pick a task, a model and a volume. Totals use today's price rules and round every run exactly as it is billed.
Total per month
$1.80
180 credits$0.0036 per image
Same volume, 6 models
Fixed prices: the amount of each run is known before it starts and rounded exactly as billed.
Add balance
What your balance buys
Choose an amount and see how far it goes. No subscription, no packages, no bonus tiers.
$25 buys about
≈ 6,944 imagesFLUX.1 [schnell]
≈ 520 imagesFLUX 1.1 Pro- ≈ 83 clips of 5 sKling v2.1
- ≈ 22,727 requests (1K in / 500 out)DeepSeek V4.1 Flash
- ≈ 1,388 × 1,000 charactersOpenAI TTS-1
- ≈ 250,000 documents of 500 tokensOpenAI text-embedding-3-small
Adding balance needs an account. Sign in
Charged in US dollars through Stripe (cards). Balance does not expire.
Real outputs
What each price point looks like
- AI-generated text
In two upbeat sentences, tell a first-time user why trying AI models on one platform is great.
Trying AI models on one platform offers a seamless and intuitive experience, allowing you to explore diverse capabilities without the hassle of switching tools. Plus, you can easily compare results and discover the most effective solutions for your needs all in one place!
GPT-4o mini≈ $0.00010Per answer, estimated from the length of prompt and answer. ![FLUX.1 [schnell]: a minimalist mountain landscape at golden hour, flat vector illustration](/_next/image?url=%2Fexamples%2Fflux-1-schnell.webp&w=3840&q=75)
AI-generated imageFLUX.1 [schnell]$0.0036/image“a minimalist mountain landscape at golden hour, flat vector illustration”![FLUX.1 [dev]: a cozy reading nook by a rainy window, warm cinematic light, shallow depth of field](/_next/image?url=%2Fexamples%2Fflux-1-dev.webp&w=3840&q=75)
AI-generated imageFLUX.1 [dev]$0.030/image“a cozy reading nook by a rainy window, warm cinematic light, shallow depth of field”Same prompt as FLUX 1.1 Pro
AI-generated imageFLUX 1.1 Pro$0.048/image“a cozy reading nook by a rainy window, warm cinematic light, shallow depth of field”Same prompt as FLUX.1 [dev]
AI-generated imageRecraft 20B SVG$0.0528/image“a clean line icon of a paper plane, minimal, single color, vector”
AI-generated imageIdeogram v3 Quality$0.108/image“a friendly robot watering a small plant, soft 3d render, pastel colors”
Outputs from real runs. Image prices are today's price per image.All image models
All prices
Model prices
| Model | Price | Example run | For $10 | Actions |
|---|---|---|---|---|
Image134 | ||||
Black Forest Labs | $0.048/image | $0.0481 image | ≈ 208 images | |
Recraft | $0.0528/image | $0.05281 image | ≈ 189 images | |
 Ideogram | $0.108/image | $0.1081 image | ≈ 92 images | |
Black Forest Labs | $0.0036/image | $0.00361 image | ≈ 2,777 images | |
 Black Forest Labs | $0.030/image | $0.0301 image | ≈ 333 images | |
Video46 | ||||
Kuaishou (Kling) | $0.060/s | $0.305 s clip | ≈ 33 clips | |
Runway | $0.060/s | $0.305 s clip | ≈ 33 clips | |
MiniMax | $0.324/video | $0.3241 video | ≈ 30 runs | |
Runway | $0.144/s | $0.725 s clip | ≈ 13 clips | |
Kling v3New Kuaishou (Kling) | $0.2688/s | $1.3445 s clip | ≈ 7 clips | |
Text & Chat21 | ||||
o3-miniNew OpenAI | $1.32/1M in$5.28/1M out | ≈ $0.00401K in / 500 out tokens | ≈ 2,500 requests | |
Google DeepMind | $1.50/1M in$12.00/1M out | ≈ $0.00751K in / 500 out tokens | ≈ 1,333 requests | |
DeepSeek | $1.584/1M in$4.752/1M out | ≈ $0.00401K in / 500 out tokens | ≈ 2,500 requests | |
Anthropic | $2.40/1M in$12.00/1M out | ≈ $0.00841K in / 500 out tokens | ≈ 1,190 requests | |
GPT-4.1New OpenAI | $2.40/1M in$9.60/1M out | ≈ $0.00721K in / 500 out tokens | ≈ 1,388 requests | |
Multimodal38 | ||||
Salesforce | ≈ $0.00030/runbilled by usage | ≈ $0.00030typical run, ≈ 1 s | ≈ 33,333 runs | |
Community | ≈ $0.0050/runbilled by usage | ≈ $0.0050typical run, ≈ 4 s | ≈ 2,000 runs | |
| ≈ $0.018/runbilled by usage | ≈ $0.018typical run, ≈ 15 s | ≈ 555 runs | ||
Community | ≈ $0.0457/runbilled by usage | ≈ $0.0457typical run, ≈ 169 s | ≈ 218 runs | |
Google DeepMind | $0.60/1M in$3.60/1M out | ≈ $0.00241K in / 500 out tokens | ≈ 4,166 requests | |
Code11 | ||||
Community | ≈ $0.0030/runbilled by usage | ≈ $0.0030typical run, ≈ 3 s | ≈ 3,333 runs | |
| ≈ $0.0065/runbilled by usage | ≈ $0.0065typical run, ≈ 6 s | ≈ 1,538 runs | ||
| ≈ $0.0074/runbilled by usage | ≈ $0.0074typical run, ≈ 6 s | ≈ 1,351 runs | ||
Community | ≈ $0.0253/runbilled by usage | ≈ $0.0253typical run, ≈ 22 s | ≈ 395 runs | |
| ≈ $0.0408/runbilled by usage | ≈ $0.0408typical run, ≈ 35 s | ≈ 245 runs | ||
Text-to-Speech14 | ||||
Community | ≈ $0.00030/runbilled by usage | ≈ $0.00030typical run, ≈ 1 s | ≈ 33,333 runs | |
Community | ≈ $0.00040/runbilled by usage | ≈ $0.00040typical run, ≈ 1 s | ≈ 25,000 runs | |
Community | ≈ $0.0024/runbilled by usage | ≈ $0.0024typical run, ≈ 2 s | ≈ 4,166 runs | |
Community | ≈ $0.0035/runbilled by usage | ≈ $0.0035typical run, ≈ 3 s | ≈ 2,857 runs | |
Haohe Liu | ≈ $0.0157/runbilled by usage | ≈ $0.0157typical run, ≈ 58 s | ≈ 636 runs | |
Audio & Music3 | ||||
Meta | ≈ $0.0529/runbilled by usage | ≈ $0.0529typical run, ≈ 31 s | ≈ 189 runs | |
Community | ≈ $0.0030/runbilled by usage | ≈ $0.0030typical run, ≈ 3 s | ≈ 3,333 runs | |
Community | ≈ $0.0069/runbilled by usage | ≈ $0.0069typical run, ≈ 6 s | ≈ 1,449 runs | |
Speech-to-Text5 | ||||
OpenAI | ≈ $0.0034/runbilled by usage | ≈ $0.0034typical run, ≈ 12 s | ≈ 2,941 runs | |
Community | ≈ $0.0056/runbilled by usage | ≈ $0.0056typical run, ≈ 5 s | ≈ 1,785 runs | |
Community | ≈ $0.0063/runbilled by usage | ≈ $0.0063typical run, ≈ 5 s | ≈ 1,587 runs | |
Community | ≈ $0.0241/runbilled by usage | ≈ $0.0241typical run, ≈ 14 s | ≈ 414 runs | |
Community | ≈ $0.156/runbilled by usage | ≈ $0.156typical run, ≈ 133 s | ≈ 64 runs | |
Embeddings2 | ||||
| $0.024/1M inbilled by usage | ≈ $0.000101K tokens | ≈ 100M tokens | ||
| $0.156/1M inbilled by usage | ≈ $0.000201K tokens | ≈ 50M tokens | ||
Top 5 of each category. Pick a category or search to see all 274.
Free
Free run and trial
There is no free tier. This is exactly what you can use without paying.
Try without an account
One free run per browser, no sign-up, on these models:
Runs for video and voice need an account.
Try it on the homepageGoogle sign-up trial
Accounts created with Google get 10 credits ($0.10) to try models.
- Usable 24 hours after sign-up
- For runs of up to 2 credits each
- At most 5 runs per 24 hours
- Enough for about 27 FLUX.1 [schnell] images
- Email and GitHub sign-ups get no trial credits
Billing
How billing works
- Fixed price
- Images, video seconds and characters are charged exactly as shown; the amount is known before the run starts.
- By usage
- Tokens and GPU time: credits are reserved when a run starts and settled to what it actually used when it ends. The rest of the reserve goes back to your balance.
- Parameters
- Prices on the website assume each model's default settings. API calls are priced for the parameters you send; fields you leave out use the provider's defaults.
- Failed runs
- Refunded to your balance automatically, unless a GPU-time run was stopped at its reserved time limit.
- Smallest charge
- 0.01 credits ($0.0001) per run.
FAQ
Frequently Asked Questions
Straight answers on what things cost and how you pay.
Billing questions: info@railwail.com
How does pricing work?
You add prepaid balance in US dollars and every run is charged from it; 1 credit = $0.01. Images, video seconds and characters have a fixed price that is shown before you start. Token and GPU-time models reserve credits when a run starts and are settled to what it actually used.
Why can a run cost more or less than the example?
Examples use each model's default settings; chat examples use the request size shown in the table. More images per run, longer clips, higher quality tiers, longer prompts or longer answers cost more. The calculator and every model page show the price for your settings before you run.
Do API calls cost the same as the website?
Both use the same price rules. The website assumes each model's default settings; API calls are priced for the parameters you send, and fields you leave out use the provider's defaults. The same settings cost the same everywhere.
Can I try models without an account?
Yes, once per browser: FLUX.1 [schnell] and DeepSeek V4.1 Flash run without sign-up from the homepage. Everything else needs an account and balance.
Is there a free tier?
There is no free tier. Accounts created with Google get 10 trial credits ($0.10), usable 24 hours after sign-up, for runs of up to 2 credits and at most 5 runs per 24 hours. Email and GitHub sign-ups get no trial credits.
Is there a subscription or a minimum spend?
No subscription and no monthly fee. The smallest top-up is $5; a single run can cost as little as $0.0001.
What payment methods do you accept?
Cards, through Stripe Checkout. You are charged in US dollars. PayPal, SEPA transfers and payment by invoice are not available.
What happens when a run fails?
Failed runs are refunded to your balance automatically. The one exception is a GPU-time run that was stopped at its reserved time limit: the time it used is charged.
Do credits expire?
No, your credits never expire. Once purchased, they remain in your account until you use them.
Can I get a refund?
Yes, unused credits can be refunded within 14 days of purchase. Contact us at info@railwail.com.
Do you offer volume discounts?
For high-volume usage, contact us at info@railwail.com to discuss custom pricing.
Pick a model and see the price before you run it
274 models with a published price. Sign up with Google for trial credits, or browse the catalog first.