Full rate card
Same source as the pricing page and the console. Unit: USD per million tokens.
| model | Input | Output | Cache-hit input | Peak input/output | Context |
|---|---|---|---|---|---|
| deepseek-flash | $0.15 | $0.6 | $0.003 | $0.3 / $1.2 | 1M |
| qwen-plus | $0.1096 | $0.274 | $0.011 | — | 1M |
| qwen-max | $0.3288 | $1.3151 | $0.0329 | — | 32K |
| qwen-flash | $0.0205 | $0.2055 | $0.0021 | — | 1M |
| auto | per routed model | per routed model | — | — | — |
Price table version 2026-09-11.1 — identical to provider rate cards, zero markup. A 10% fee applies to top-ups.
Upstream and data residency
| model | Upstream / jurisdiction |
|---|---|
| deepseek-flash | DeepSeek first-party API (api.deepseek.com) |
| qwen-plus | Alibaba Cloud Model Studio (Mainland China node) |
| qwen-max | Alibaba Cloud Model Studio (Mainland China node) |
| qwen-flash | Alibaba Cloud Model Studio (Mainland China node) |