Estimate your LLM API cost before you commit
Pick a model, enter your monthly token volume, and set a cache-hit rate. The calculator uses our real published multipliers on a 0.4 CNY per 1M token base — no sign-up required.
How this calculator differs from guessing
Most cost estimates are back-of-envelope. This one runs on the same per-model multipliers the gateway actually bills.
| This calculator | A rough estimate | |
|---|---|---|
| Rate source | Live published per-model multipliers | Memory or a blog post |
| Cache modeling | Adjustable cache-hit rate from the 0.1x floor | Usually ignored |
| Currency | CNY, matching how you are billed | Mixed USD/CNY |
| Model coverage | Every text model the gateway routes | One or two models |
Who this is for
Useful if you
- Want a defensible cost estimate before moving traffic to a model
- Need to compare a 0.1x model against a frontier model on the same volume
- Want to see how a higher cache-hit rate lowers your effective rate
- Are budgeting a monthly LLM spend in CNY
Less useful if you
- Need USD figures — this calculator works in CNY
- Bill per image or per call rather than per token (those rows are excluded)
- Want a guaranteed quote — actual usage depends on your prompts and outputs
- Need provider-native pricing for a model the gateway does not route
How to read the result
The math is simple and transparent, but estimates are estimates.
- Cost = effective rate × your token volume. The effective rate blends the normal and cached rates by your cache-hit percentage.
- The cached rate uses the documented 0.1x floor of the model multiplier; real cache savings depend on your prompt reuse.
- Token counts are yours to estimate — input plus output. Real usage varies with prompt and response length.
- Image and per-call models are excluded because they are not billed per token.
The rule behind the numbers
Nothing here is hidden. Every model's rate is its multiplier times the base.
This is an estimate based on published rates and your inputs, not a quote. Actual cost depends on real token usage and cache behavior in your workload.
Estimate only, based on published multipliers and your inputs. Image and per-call models are excluded.
Pricing calculator — FAQ
Where do the rates come from?
From the gateway's published per-model multipliers applied to a 0.4 CNY per 1M token base. The same rule the pricing page lists.
What does the cache-hit rate do?
Cached input bills from the 0.1x floor of the model's multiplier. Raising the cache-hit percentage lowers your blended effective rate.
Is the result a guaranteed price?
No. It is an estimate. Your real cost depends on actual input and output token counts and how much of your input is cached.
Why is everything in CNY?
Because billing and prepaid credits are denominated in CNY. The calculator matches how you are actually charged.
Happy with the estimate?
Buy a key and start — prepaid credits, no expiry.