Skip to main content
The VLM Run Gateway meters every request. Every served model bills per token. The rates here come straight from the live model catalog (GET /v1/models/{model_id}), so they always match what you are actually charged. Two signals let you wire up cost tracking today:
  • Per-model rate card. Each model exposes a pricing object with USD-per-1M rates for prompt, completion, input_cache_read, input_cache_write, and image.
  • Per-request cost. usage.cost on a chat completion, streaming or not, is the actual metered charge for that request, in USD rounded to six decimals, the resolution the ledger bills at. Embeddings and transcriptions report token counts only, so multiply them by the rate below.
The rates below are the currently published rates and may change. Always read pricing from the model catalog (or usage.cost on a chat completion) as the source of truth.

​
How billing works

  • Token rates are USD per 1M tokens. Input is prompt tokens, Cached input is input_cache_read, and Output is completion tokens. A request pays prompt on the uncached prompt tokens, input_cache_read on the cached ones, and completion on the output.
  • Documents bill per page. The Gateway rasterizes a PDF and runs one inference per page. It sums the per-page token counts, so the charge is linear in page count.
  • input_cache_write is $0.00 across all models today. The image rate is published for reference. No served model bills on it today, because image and audio inputs already count as prompt tokens.
  • Neither method (ocr, markdown, detect, parse_layout, chat) nor response_format (text or JSON) changes the rate. You pay the rate of whichever model serves the request.
  • Non-generative vision models emit no text. They report a flat prompt-token count per image instead, so the token rate gives a fixed price per image.

​
What 1M tokens buys

A 1M-token budget covers roughly the following:
  • ~3,000 image captions at ~200 input + 130 output tokens / image
  • Markdown from a ~2K-page slide deck with figures at ~500 output tokens / page
  • ~4,000 visual questions at ~200 input + 50 output tokens / question
  • ~2,000 text passages embedded at ~512 tokens / passage
  • Markdown from ~1K pages of legal docs with dense text at ~1,000 output tokens / page

​
Chat and document models

Prices in USD per 1M tokens

​
Pose estimation

usyd-community/vitpose-plus-large emits no text. It reports a flat 256 prompt tokens per image and no output tokens, so one image costs about $0.001, or about 1,000 images per dollar.

Prices in USD per 1M tokens

​
Routed models

The routed models also bill per token, at the provider’s listed rates. They carry the paid access tier, so an organization needs an active subscription or a funded balance to call them.

Prices in USD per 1M tokens

​
Embeddings

Text, image, and video inputs all bill as prompt tokens. There is no output charge.

Prices in USD per 1M tokens

​
Transcription

Transcription bills the audio it ingests as prompt tokens. There is no separate output charge.

Prices in USD per 1M tokens

​
How to get model pricing

cURL

Get Model

Full response schema, including pricing.

Rate Limits

Per-tier request quotas.