Published 2026-08-16 · 赵予安
Direct answer
DeepSeek V4 PRO CN is listed at 0.239 / 0.479 per million tokens for Chinese workloads. Context window and CacheSafety stay unpublished. This guide is written for Singapore product and platform teams comparing model quality, spend, routing policy, and production rollout risk.
What shipped
Deepseek V4 PRO CN is listed in the NextModel gateway catalog as deepseek/deepseek-v4-pro-cn. The public slug is deepseek--deepseek-v4-pro-cn. The provider is DeepSeek. Streaming is the listed capability. The latency band is 1000 to 3000 milliseconds.
Pricing
Input price is $0.239 per 1 million tokens. Output price is $0.479 per 1 million tokens. The context window is unpublished in the catalog snapshot. Related catalog entries show the surrounding tier. Deepseek V3.2 CN is $0.042 per 1 million input tokens and $0.064 per 1 million output tokens. Deepseek V4 Flash CN is $0.020 per 1 million input tokens and $0.041 per 1 million output tokens.
Who this is for
The catalog lists Chinese as the use case. This model is for developers who need Chinese-language responses with streaming support. It has a higher price point than the V3.2 CN and V4 Flash CN tiers.
What is still unverified
The release date is unpublished in this catalog snapshot. Independent third-party quality scores are not in these notes. The context window is unpublished when the catalog value is zero. The NextModel bench table has no first-party CacheSafety row for this exact model. Safe Hit, Bad Hit, and trap rates have no published values.