Pay as you go, or by the month
Start with per-token billing and a balance you top up. Move to a monthly plan when you want predictable spend, higher priority and more concurrency.
Per-token rates
| Model | Input / 1M | Cache hit / 1M | Output / 1M | Context |
|---|---|---|---|---|
| DeepSeek V4 Flash | $0.14 | $0.003 | $0.28 | 1M context |
| DeepSeek V4 Pro | $0.60 | $0.005 | $1.20 | 1M context |
Cache hit = repeated context the model already cached — billed up to ~98% cheaper, just like the upstream. Prepaid balance · top up from $5 · billed to the token · DeepSeek V4 today, more models soon.
More keys, more models
Basic
No card required
- 1 API key
- Flash model only
- Up to 128K context
- Pay-as-you-go balance
- OpenAI-compatible API
- Streaming + cached-input savings
Pro
Fee converts to balance
- 10 API keys
- Flash + Pro models
- Up to 256K context
- Fee converts to balance
- OpenAI-compatible API
- Streaming + cached-input savings
Max
Fee converts to balance
- 50 API keys
- Flash + Pro models
- Up to 1M context
- Fee converts to balance
- OpenAI-compatible API
- Streaming + cached-input savings
Paid plans are billed monthly via Paddle and the fee is credited to your balance — so you only pay for the membership perks (more keys, all models), not extra on top of usage.
Handled by Paddle
Payments and sales tax are processed by Paddle.com, our Merchant of Record. See our refund policy.
Prices will drop
Today we relay the official DeepSeek API. As we bring our own compute online, savings get passed back to lower rates.
No infinite plans
Every plan has clear daily and concurrency limits, so heavy users never degrade everyone else.