Reusable prompt
Cache assumptions
Estimate savings on repeated system instructions and shared context.
Prompt Cache SavingsCaching
Estimate how much reusable system prompts can save across OpenAI, Anthropic, and Google.
Current provider pricingNothing uploadedFree
Verified reference
Calculation reference
Reference values · verified Oct 2026
| Provider | Standard input | Cached read | Approx. discount |
|---|---|---|---|
| OpenAI · GPT-6.1 Sol | $2.00 / 1M | $0.10 / 1M | ≈95% |
| Anthropic · Sonnet 5.5 | $2.00 / 1M | $0.10 / 1M | ≈95% |
| Google · Gemini 4 Argon | $2.00 / 1M | $0.10 / 1M | ≈95% |
Clear answers
Frequently asked questions
Caching requires repeated prompt prefixes that meet each provider’s minimum length and retention rules.
Keep calculating