Conversation / Quote
DeepSeek-V4-Pro API Price
DeepSeek · deepseek-v4-pro
| Billing | CNY · 1M tokens |
|---|---|
| Context Window | 1000K tokens |
| Max Output | 393K tokens |
| Modalities | Text |
| Deep Thinking | Supported |
| Verified Date | 2026-10-12 |
| Official pricing page | api-docs.deepseek.com/zh-cn/quick_start/pricing ↗ |
Source text: the note below is the model's official text, kept in its original Chinese and not translated.
MODEL VERSION: DeepSeek-V4-Pro-0813；CONTEXT LENGTH 1M，MAX OUTPUT 384K；Vision Not supported；Off-peak rates are half of the peak rates；顶层 input/output/cacheRead 取 PEAK（标准）价，分时价见 timedPrices
Report a pricing issue
Found inaccurate pricing or model information? We will review it.
Time-Based Pricing
| Tier | Price | Description |
|---|---|---|
| Off-peak hours | in ¥4.5 / out ¥13.5 /1M tokens | — |
| Peak hours | in ¥9 / out ¥27 /1M tokens | — |
Monthly Cost Estimate
30 days · no cache hits · list price
Free quotas, tiered pricing, batch discounts, and time-based pricing are not included; refer to official bills.
DeepSeek Related Models
- DeepSeek-V4.1-Flash ¥1 / ¥4
- DeepSeek-V4-Flash (legacy) ¥1 / ¥4
- DeepSeek-V4-Flash(v0731) ¥3 / ¥9
- DeepSeek-V4-Flash-Vision-Exp (legacy) ¥1 / ¥4
- DeepSeek-R1(Legacy) ¥4 / ¥16
- DeepSeek-R1-0528(Legacy) ¥4 / ¥16
- DeepSeek-V3(Legacy) ¥2 / ¥8
- DeepSeek-V3-0324(Legacy/Open-source) ¥2 / ¥8
- DeepSeek-V3.2(Legacy) ¥2 / ¥3
- DeepSeek-R1-Distill-Qwen-7B(Open-source) ¥0.5 / ¥1
- DeepSeek-R1-Distill-Qwen-14B(Open-source) ¥1 / ¥3
- DeepSeek-R1-Distill-Qwen-32B(Open-source) ¥2 / ¥6
- DeepSeek-R1-Distill-Llama-70B(Open-source) Free
- DeepSeek-V3.1(Legacy) ¥4 / ¥12
- DeepSeek-OCR(open-source hosted) ¥0.216 / ¥0.216
FAQ
How much does the DeepSeek-V4-Pro API cost?
Input ¥4.5/1M tokens, output ¥13.5/1M tokens
How large is the DeepSeek-V4-Pro context window?
1000K tokens, up to 393K tokens of output. Handles long documents, multi-turn conversations and large-scale inputs.
Which input modalities does DeepSeek-V4-Pro support?
Text. Supports deep thinking / reasoning.
What is the approximate monthly cost of DeepSeek-V4-Pro?
Depends on volume: light use (1M input + 300K output per day) ≈ ¥256.5/month; medium (5M + 1.5M) ≈ ¥1,282.5/month; heavy agents (10M + 3M) ≈ ¥2,565/month. Cache hits, batch discounts and tiered pricing can lower this further.
How does DeepSeek-V4-Pro compare with peers on value?
DeepSeek DeepSeek-V4-Pro is priced low among comparable models; compare against GPT-4o, Claude 3.5 and Gemini 2.5. See the Pricemon market table for a full comparison.
What use cases suit DeepSeek-V4-Pro?
Good for: chat systems, customer support, content generation, coding assistance, data analysis.The long context supports deep-analysis scenarios.
Does DeepSeek-V4-Pro have a free tier?
Check DeepSeek's official pricing page for the latest free quota, trial policy and tiered pricing. Policies differ by vendor and change over time.