Prompt Caching Savings Calculator
See how much prompt caching cuts your input cost.
With the default inputs, 72.0% cheaper input — Save ~$1,620.00/month on input tokens. Enter your own numbers below to recompute instantly; the full step-by-step math is shown under the worked example.
Results update automatically as you type.
- Input cost without cache
- $0.0150
- First request (writes cache)(one-time cache-write premium)
- $0.0180
- Cache-write premium (first hit)
- $0.0138
- Cached request (cache hit)
- $0.00420
- Savings per cached request
- $0.0108
- Savings per month
- $1,620.00
Prompt caching bills repeated input tokens at a fraction of the normal rate. Set how much of your input is cacheable and your volume to see the monthly savings.
How this is calculated
Savings = cacheable input tokens × requests × (normal input price − cached price). Cache hits are billed at roughly 10% of the normal input rate on most providers, so a large, stable prefix (system prompt, tool definitions, retrieved context) reused across requests is where caching pays off.
Is this a good result? What to do next
The more of your input is stable and reused, the bigger the win. Caching a large system prompt across high volume can cut the cost of the cached share by roughly 90%. If little of your input repeats, caching does almost nothing.
Typical planning ranges
- Cache read price
- ~10% of normal input
- Best case
- stable prefix + high reuse
- Low benefit
- unique prompts every call
Ranges are typical planning figures to sanity-check your result, not authoritative benchmarks. Your numbers will vary with use case, volume, and vendor.
How to improve this number
- Maximize the cacheable, byte-stable prefix.
- Reuse the cached prefix within its TTL to keep hit rates high.
- Put the volatile part of the prompt after the cached prefix.
Common mistakes
- Caching content that changes each call (never hits).
- Ignoring cache TTL and minimum-size requirements.
When to use a different approach
For the fixed system prompt specifically, use the system prompt cost calculator. For async discounts, use the batch API savings calculator.
Worked example (defaults)
With the default inputs above, here is the result:
- Input cost without cache
- $0.0150
- First request (writes cache)(one-time cache-write premium)
- $0.0180
- Cache-write premium (first hit)
- $0.0138
- Cached request (cache hit)
- $0.00420
- Savings per cached request
- $0.0108
- Savings per month
- $1,620.00
Sources & references
Frequently asked questions
What can be cached?+
Stable prefixes like system prompts, tool definitions and retrieved context. Cache hits are billed at roughly 10% of the normal input rate on most providers.
Related calculators
Answering a real question?
This calculator powers these problem-solving guides: