Model Plaza
The Model Plaza lists every model Rabah offers, with its prices and the name to call it by.
- Signed out, it shows public lines at their base prices.
- Signed in, it also shows exclusive lines granted to you, and prices use your personal rate.
Finding a model
Section titled “Finding a model”- Filter by platform at the top: Anthropic (Claude), OpenAI (GPT), Gemini, Grok and others.
- Or type part of a model name into the search box, e.g.
sonnetorgpt-5.
Copying the name to call
Section titled “Copying the name to call”The model in your request is the name shown here. Click Copy model name:
- On the model’s row, it copies the bare name, e.g.
claude-sonnet-4-6, which uses the model’s default line. - On a line inside the expanded row, it copies
model:line-code, e.g.claude-sonnet-4-6:pro, which always uses that line.
A model with several lines shows “{n} lines” on its row; expand it to see and compare them.
Reading the prices
Section titled “Reading the prices”- Token-priced models: prices are US dollars per million tokens. An input price of
$3means $3 for every million input tokens. - Models priced per request, image or second (some image and video models) are labelled “Per request”, “Per image” or “Per second”.
Columns
Section titled “Columns”| Column | Meaning |
|---|---|
| You pay | The main price you pay, with the rate applied: the input price for token-priced models, the unit price for per-request or per-image models |
| Output / secondary | The output price, plus cache write, cache read and other prices |
| Official input, Official output | The model maker’s list prices, for comparison |
| Rate | How your price compares with the official one |
The main table shows the default line’s prices. Expand a model to compare its lines.
You pay = official price × rate.
- Each line has its own rate. A rate of
0.5is half the official price. - If you have a personal rate, it replaces the line’s rate; the two don’t stack. Signed in, the plaza shows your personal rate.
Peak hours
Section titled “Peak hours”Some lines have peak hours, when prices are multiplied by a peak rate. The plaza shows the window, the multiplier and the time zone next to the line, e.g. 14:00-18:00 ×1.5; “Peak” in the table is the price during those hours.
Other prices
Section titled “Other prices”| Name | Meaning |
|---|---|
| Cache write 5m / 1h | Writing content to the prompt cache, for a 5-minute or 1-hour lifetime |
| Cache read | Reading content back from the cache, usually far cheaper than normal input |
| Image input | The price of image input when it is billed by tokens |
| Context tier | Some models price by how long a request’s context is; long contexts cost more |
| Channel schedule | Some models cost different amounts at different times of day; follow the hours and time zone shown |
What you are actually charged is in the per-request details under Usage in the Hub.