Shadow SoftwareBuy credit
Shadow Inference

One OpenAI-compatible endpoint. Priced per token.

Routes each request to a model with enough context window and output room. Change your base URL and keep your client. The models and rates below are exactly what your account can use.

POST https://inference.shadowsoftware.com/v1/chat/completions  ·  model: "auto" or a model id from the table

Prices are USD per million tokens, billed on actual input and output tokens. Context is the model's window; max output is the most it can generate in one reply (— means no cap is published). Struck-through figures are the regular price before the current discount. Data as of 2026-11-18. Need something not listed? Ask us.