Transparent pricing with no hidden fees. Pay as you go.
deepseek-v4-prodeepseek-v4-pro| Model | Spec | PoYo price | Official price | You save |
|---|---|---|---|---|
| deepseek-v4-pro | Input tokens | $0.348/1M tokens 68.4 credits /1M tokens | $0.435/1M tokens Official | 20% |
Output tokens | $0.696/1M tokens 136.8 credits /1M tokens | $0.870/1M tokens Official | 20% |
* Actual fees are based on the final output.
Complete guide to using DeepSeek V4 Pro API
deepseek-v4-pro. DeepSeek documents it as a stronger reasoning, coding, and agent workflow model with 1.6T total parameters and 49B active parameters.DeepSeek V4 Preview is officially available with 1M context and separate Flash and Pro profiles.
01
DeepSeek V4 uses a mixture-of-experts profile. Official docs list DeepSeek V4 Pro at 1.6T total / 49B active, giving teams a clear separation between the efficient Flash route and the stronger Pro route.
02
DeepSeek positions V4 Pro as open-source SOTA for agentic coding benchmarks, strong in Math/STEM/Coding, and competitive with top closed-source models for reasoning-heavy work.
03
The official DeepSeek release notes state that 1M context is the default across official DeepSeek services. This is useful for long chats, repository-scale coding context, large documents, and multi-step agent memory.
04
DeepSeek documents dual modes for the V4 API family. Use non-thinking for faster direct answers and thinking mode for tasks where explicit reasoning effort improves quality.
Use Pro for repository-scale reasoning, code review, migrations, and complex debugging.
Run harder reasoning prompts where the stronger Pro profile is worth the extra credits.
Analyze policies, technical specs, and research packs with longer context and stronger synthesis.
Pair Flash and Pro so routine traffic stays cheap while difficult requests move to Pro.
Both Model IDs come from the same DeepSeek V4 Preview API family. Choose Flash for efficient traffic and Pro for harder reasoning, coding, and agent work.
| Feature | DeepSeek V4 Flash | DeepSeek V4 Pro |
|---|---|---|
| Model ID | deepseek-v4-flash | deepseek-v4-pro |
| Total / active parameters | 284B / 13B | 1.6T / 49B |
| Positioning | Fast, efficient, economical | Stronger reasoning, coding, agents |
| Context window | 1M context | 1M context |
| Modes | Thinking / Non-Thinking | Thinking / Non-Thinking |
| Pricing | 22.8 input / 45.6 output credits per 1M tokens | 68.4 input / 136.8 output credits per 1M tokens |
Sources: DeepSeek official V4 Preview release. Official source: https://api-docs.deepseek.com/news/news260424
Create a PoYo API key
Get API Key
Set the model ID
Use deepseek-v4-pro as the model ID in your request.
Open API docs
View API docs
| Model | Input | Output |
|---|---|---|
| DeepSeek V4 Pro | $0.348 / 1M tokens 68.4 credits | $0.696 / 1M tokens 136.8 credits |
deepseek-v4-pro.