Input tokens
$0.30per 1 million tokens
Uncached input only. Cache reads use their separate rate.
DEEPSEEK V4.1 FLASH 1017 API PRICING
Prepaid credits.
The same rates for every pack.
Choose $10, $20, or $50–$200 in DeepSeek V4.1 Flash 1017 API credits. Input, output, and cache reads share one USD balance.
No subscription or automatic renewal. Every pack uses the same rates.
Output tokens
$1.20per 1 million tokens
Includes reasoning tokens reported as output.
Cache read tokens
$0.0024per 1 million tokens
Explore
$10 USD
Try a few prompts and connect your first app.
$10 in API credits
If used entirely for one token type:
- Input tokens
- 33.33M
- Output tokens
- 8.33M
- Cache read tokens
- 4,166.67M
Each figure shows what the full balance could cover on its own. A mix of token types uses the same balance.
Choose $10 pack →Build
$20 USD
Keep building and testing as your project grows.
$20 in API credits
If used entirely for one token type:
- Input tokens
- 66.67M
- Output tokens
- 16.67M
- Cache read tokens
- 8,333.33M
Each figure shows what the full balance could cover on its own. A mix of token types uses the same balance.
Choose $20 pack →Scale
$50 USD
Add more credit for regular use or larger projects.
$50 in API credits
If used entirely for one token type:
- Input tokens
- 166.67M
- Output tokens
- 41.67M
- Cache read tokens
- 20,833.33M
Each figure shows what the full balance could cover on its own. A mix of token types uses the same balance.
Choose $50 pack →
ESTIMATE YOUR USAGE
Estimate your API costs.
Enter millions of tokens. Input here means uncached input; enter cached input separately so it is counted once.
Estimated usage cost: $1.50
Your bill is based on the token usage returned with each request. New requests and follow-up prompts add to the total.
How your balance works
One balance for your account
Each confirmed payment adds its USD amount to your balance. Playground and API requests use that same balance.
Pay for the tokens you use
Uncached input, cache-read input, and output are billed separately. Reasoning tokens already included in output are not charged again.
How temporary holds work
We place a temporary hold before sending a paid request. When usage is confirmed, we charge the actual cost and release the rest. If the result cannot be confirmed, the hold stays in place while we review it.
View payments and usage
Check orders, refresh a pending payment, review usage, or request a refund from your account. Credits appear after the payment provider confirms your payment.
Pricing questions
Do larger packs have lower token rates?
No. Every pack uses the same input, output, and cache-read rates. Larger packs add more USD credit, without a volume discount.
How much does a mixed request cost?
For example, 1M uncached input tokens plus 1M output tokens costs $1.50. If the input is instead 1M cache-read tokens, the total is $1.2024. A cache hit depends on the upstream receipt and is not guaranteed.
Which payment methods can I use?
Choose Stripe or PayPal from the options available at checkout. Stripe shows the payment methods supported for your purchase. If checkout is unavailable, you can still view existing orders and request help with a payment.
Is a free trial included?
The free trial includes 20,000 input and output tokens, valid for seven days after claim. Cached input counts toward the same allowance. Check your verified account to see whether you can claim it. Trial requests never use paid credit automatically.
Can I request a refund?
Submit an order-specific request from Credits and payments. Requests are reviewed before any refund. Read our refund information for eligibility and the review process.