Model Pricing
DDTokens charges for actual token usage. The final cost depends on the model, provider, cache hits, and your account benefit tier.
How Pricing Is Calculated
A request usually has three cost components:
- Input tokens: user messages, system prompts, and context.
- Cached tokens: some models charge less when prompt caching is used.
- Output tokens: content generated by the model.
The final cost can be understood as:
Request cost = standard model price x actual token usage x benefit discountYour actual discount is determined by your account tier and the price shown on the pricing page.
Example Prices
The prices below only illustrate the billing fields. Official prices are shown in the dashboard or on the live pricing page.
| Model | Input price | Output price | Unit |
|---|---|---|---|
| GPT-4o | $2.50 | $10.00 | per million tokens |
| GPT-4o Mini | $0.15 | $0.60 | per million tokens |
| Claude 3.5 Sonnet | $3.00 | $15.00 | per million tokens |
| Gemini 1.5 Flash | $0.075 | $0.30 | per million tokens |
User Tiers and Discounts
Your user tier may affect:
- Model price discounts.
- Request rates.
- Free-model capacity.
- Support for high-concurrency or enterprise use cases.
See Benefit Tiers to learn how DDTokens organizes tiers and benefits.
Why the Same Model May Have Different Prices
The same model name may be supplied by different upstream providers. Provider costs, capacity, response formats, and availability can differ, so prices may differ as well.
If you do not select a provider, DDTokens uses default routing. This is the best option for most use cases. If you need to pin a provider or stabilize provider-specific behavior, read Providers & Routing.
View Usage and Costs
Sign in to the dashboard to view request logs and billing details. When investigating costs, focus on:
- Requested model and actual processing model.
- Input, cached, and output tokens.
- Cost before discount and actual cost.
- Request duration and upstream duration.