Add credit with crypto and pay only for the tokens you use. There is no monthly plan, no seat fee and no email requirement. The minimum top-up is $20, credits never expire, and failed requests are not charged.
Prices below are USD per 1M tokens. Enter an input and output size to estimate one request. The estimate excludes cached-input discounts and is not an invoice.
Example: Magnum V4 72B at $12 input / $20 output per 1M tokens.
These are the models most customers use for creative writing, roleplay, research and agent workflows. Every enabled model appears in the full rate sheet with its live price and context window.
| Model | Type | Context | Input | Output |
|---|---|---|---|---|
| Magnum V4 72B magnum-v4-72b |
uncensored | 32.8K | $12 | $20 |
| DeepSeek V4.1 Flash 1017 Uncensored deepseek-v4.1-flash-uncensored |
uncensored | 1.05M | $0.6 | $2.4 |
| Weaver (Alpha) weaver |
uncensored | 8.0K | $2 | $3 |
| Cydonia 24B V4.1 cydonia-24b-v4.1 |
uncensored | 131.1K | $1.2 | $2 |
| Skyfall 36B V2 skyfall-36b-v2 |
uncensored | 32.8K | $2.2 | $3.2 |
| UnslopNemo 12B unslopnemo-12b |
uncensored | 1.02M | $1.6 | $1.6 |
| ReMM SLERP 13B remm-slerp-l2-13b |
uncensored | 6.1K | $1.8 | $2.6 |
| GLM-5.3 Flash Uncensored glm-5.3-flash-uncensored |
uncensored | 1.05M | $0.165 | $0.618 |
Charge = uncached input tokens x input rate + cached input tokens x cached rate + output tokens x output rate. Token counts come from the model provider. Before each call we hold the maximum possible cost against your balance and release the difference after the response. A request that fails upstream is not charged.
Prices cover model inference only. Your own hosting, network costs, application storage and any third-party tooling are separate. Crypto network fees and exchange spreads are borne by the sender.
Per token. You pay the input rate for uncached prompt tokens, the cached rate where a provider supports prompt caching, and the output rate for generated tokens. Token counts come from the model provider.
No. There is no subscription, seat fee or minimum monthly spend. You add credit and it is spent only when you make requests.
The minimum is $20. The maximum per invoice is $5000. Credit is the USD value actually received after network confirmation.
No. Credits stay on the account until they are spent. Top-ups do not renew automatically.
No. A request that fails before a response is completed is not charged. The maximum estimated cost is held before the call and the unused amount is released afterwards.
The margin is on the uncensored and abliterated models. Frontier models are passed through at the upstream list price so one account can cover the whole workload.