The DeepSeek V4 family from its maker's own API
Hangzhou DeepSeek Artificial Intelligence Co., Ltd. makes the DeepSeek models, and this route is its own API rather than a third-party host of the open weights. On Opper it carries V4 Flash — both the 0731 checkpoint and the floating latest tag — V4 Pro 0813, and the experimental vision build of V4 Flash, which no other provider serves. All of them run at 1M context with a 384K output ceiling. Hosted in mainland China. The route to pick when you want the checkpoints straight from the maker, or need the vision variant.
Every model we route through DeepSeek. Compare residency, ZDR, training posture, and price at a glance — full data-handling detail per route below.
| Model | Region | Zero data retention | Training | Context | Input | Output |
|---|---|---|---|---|---|---|
| Multi | Not by default | Yes | 1M | $1.32 | $3.96 | |
| Multi | Not by default | Yes | 1M | $0.44 | $1.32 | |
| Multi | Not by default | Yes | 1M | $0.44 | $1.32 | |
| Multi | Not by default | Yes | 1M | $0.44 | $1.32 |
DeepSeek hosts on 1 route. Each route has its own privacy posture, residency, and GDPR terms. Postures are maintained by Opper with a last-verification timestamp.
Zero data retention is not on by default — provider logging. Trained on customer data by default. CN.
