LICENSEWARE
Mistral AI Metric

Cached input (/M tokens)

Catalog row in Vendor License Metrics · Cited

Aliases
cached input tokens; prompt caching
Definition
“Cached input tokens reduce input cost by up to 90% for repeated prompts.” (Mistral pricing FAQ.) Mistral Docs Pricing lists a Cached input column per model.
Counting rules
Applies to input tokens served from cache on repeated prompts; priced per million tokens, listed per model at one tenth of the input rate in the Docs pricing table (for example Mistral Large 3: USD 0.5 input, USD 0.05 cached input).
Floors / minimums
None published.
Applies to products
  • Mistral Large 3
  • Mistral Medium 3.5
  • Mistral Small 4
  • Ministral 3 14B
  • Ministral 3 8B
  • Ministral 3 3B
Related metrics
Product category
Esc