Formal Engines Inference

The same models, thirty to fifty per cent cheaper.

Point any OpenAI or Anthropic client at one base URL. Prepaid credit, per-token billing, no minimum and no contract.

Our price against the OpenRouter list price, per million tokens
Model Output, list You pay
Loading rates

Per million output tokens. Input costs less. Cached input costs a tenth.

30%from the first token
40%from $2,000 a month
45%from $10,000 a month
50%from $20,000 a month

Discounts apply automatically as your spend crosses each threshold.

Sign in or create an account

No password. We send a link that works once and expires in fifteen minutes.


Once you are in

  1. Add credit

    Pay in USDC or any of 350 other coins. Credit lands in minutes.

  2. Copy your key

    Shown once. Cap what each key can spend in a day.

  3. Change one line

    https://formalengines.com/inference/v1

Full rate card · Talk to us