deepseek v4
deepseek-v4-flash
Input $0.14 / 1M tokens, output $0.28 / 1M tokens. Capabilities: text.
Open model calculatorProvider cost planner
DeepSeek API Cost Calculator
DeepSeek API models for chat, reasoning, coding, and token cost planning.
Estimate monthly AI API costs for DeepSeek models.
Calculator
Estimate DeepSeek API cost
DeepSeek is preselected with deepseek-v4-flash as the starting model.
Model selection
Choose the provider and model you want to estimate.
DeepSeekdeepseek-v4-flash
Input price
$0.14 / 1M tokens
Output price
$0.28 / 1M tokens
Last updated: 2026-07-09View pricing source
Usage assumptions
Estimate traffic and token usage for an average request.
Estimated results
Run the calculator to see usage volume and projected cost.
Add your usage assumptions, then calculate to preview estimated AI API spend.
Pricing summary
DeepSeek model pricing overview
These cards summarize the provider data currently used by the calculator.
Models tracked
2
Cheapest input
$0.14 / 1M tokens
Cheapest output
$0.28 / 1M tokens
Cached input pricing
2 models
Context window range
Not listed
Last updated
2026-07-09
Model pricing
DeepSeek models
Compare provider models with the same pricing unit before opening a model-specific calculator.
Swipe sideways to see all columns.
| Model | Input price | Output price | Cached input | Context window | Capabilities | Action |
|---|---|---|---|---|---|---|
deepseek-v4-flash deepseek v4 | $0.14 / 1M tokens | $0.28 / 1M tokens | $0.0028 / 1M tokens | Not listed | text | Open model |
deepseek-v4-pro deepseek v4 | $0.435 / 1M tokens | $0.87 / 1M tokens | $0.003625 / 1M tokens | Not listed | text | Open model |
Related models
Open a DeepSeek model page
Each model page keeps the calculator preselected to that provider and model.
deepseek v4
deepseek-v4-pro
Input $0.435 / 1M tokens, output $0.87 / 1M tokens. Capabilities: text.
Open model calculatorRelated providers
Compare with other providers
Open another provider page to run the same calculator assumptions against different model options.
OpenAI provider page
OpenAI API models for chat, automation, content generation, and product AI features. 9 models tracked.
Anthropic provider page
Anthropic Claude API models for assistants, document workflows, and AI product features. 14 models tracked.
Gemini provider page
Google Gemini API models for chat, multimodal workflows, automation, and SaaS features. 10 models tracked.
Next steps
Related calculators and comparisons
Continue from this provider into focused calculators or comparison pages that match common evaluation paths.
OpenAI vs Anthropic Pricing
Compare OpenAI and Anthropic pricing assumptions, then run the same usage scenario through the calculator before choosing a provider.
GPT-4o Mini vs Claude Haiku Cost
Compare the closest available OpenAI and Anthropic small-model pricing data, then calculate your own usage scenario.
AI Chatbot Cost Calculator
Estimate product chat and assistant usage from users, messages, and tokens.
Customer Support AI Cost Calculator
Model support automation costs from tickets, replies, and token usage.
SaaS AI Cost Calculator
Estimate SaaS AI feature cost per user and margin impact.
FAQ
DeepSeek cost planning questions
Short answers for using this provider calculator.
How is DeepSeek API cost calculated?
The calculator multiplies input, output, and cached input tokens by the selected model pricing, then scales the result by request volume.
Which DeepSeek model should I choose?
Start with the model that matches your context, output, and capability needs, then compare monthly cost with the same usage assumptions.
Does this estimate include cached input pricing?
Cached input pricing is shown when the selected model has cached token pricing in the data layer.
How accurate are these estimates?
They are planning estimates based on current local pricing data and your inputs. Check the provider source before making budget decisions.
Can I compare DeepSeek with other providers?
Yes. Use the providers page or related provider links to run the same scenario against other provider models.