1M context, reasoning tiers, tool calling. Prepaid credits, metered per token.
| Model | Input | Cached input | Output |
|---|---|---|---|
| abliterated | $0.2 | $0.05 | $0.9 |
| standard | $0.111 | $0.007 | $0.165 |
Per 1M tokens. Pass the model id in the
model field. Cached input is metered separately, so repeated context
across turns costs a fraction of fresh input.
Each key allows 8 concurrent requests and 600 requests per minute. All customers share one 4xB200 replica, so concurrency is metered to keep batches fair. Open additional keys for more parallelism. Minimum top-up $20; credits never expire.
Add credits to activate your account and issue keys.
Already have an account? Dashboard