Start
How a request is priced
A request is priced in tokens: input, output, and cache are separate per-million rates. Minimum $0.005 per call, even on a tiny reply.
Six models, six cards on the pricing page. Here is how billing works, not the full price list.
How it is counted
The client sends Chat Completions to https://uncensored.monster/v1. The bot balance is charged: input + output + cache hit for the model id. Empty model routes to GLM (z-ai/glm-5.3-flash-uncensored).
GLM example: input $0.39, output $1.58, cache $0.16 per 1M tokens. Other ids and figures live only on /pricing/.
Minimum and zero
Minimum $0.005 per request. The API has zero free calls — unlike bot chat (1 request per month, 2 with Premium).
The Telegram model picker does not change API billing. Which id to pick is in the six-model comparison.
Packs
Top up in @uncensobot: $1, $5, $10, $25, $50, $100. Crypto or xRocket USDT. No payment on the website.
Why PAYG beats a self-hosted GPU on RunPod (~$40/hour idle) is in PAYG from one dollar. Setup: how to start, protocol: docs.
FAQ
Why was I charged $0.005 on a short reply?
That is the per-request minimum. Above that, billing follows the model card per token.
Where is the full price list for all six models?
On the pricing page. We do not clone every table in the blog.
Does 402 mean no balance?
Yes. Buy a pack in the bot. Key and Base URL stay the same.