Supports two reasoning effort levels: high (faster, for multi-step agent tasks) and xhigh (deeper thinking, for math and complex reasoning). If you need to control reasoning depth, ask for high or xhigh mode explicitly.
Schema documentationNo tools enabled.
| Type | Credits | Units |
|---|---|---|
| Input | 3.99 | Credits per 1k tokens |
| Output | 33.25 | Credits per 1k tokens |
| Cache hit | 0.80 | Credits per 1k tokens |
No variants available for this model.
