Methodology: Prompt Cache ROI

Model the payback of prompt caching given your hit rate.

Citations last refreshed: 2026-10-21 Pricing snapshot: 2026-10-26 ← Back to calculator

How we keep this honest

Every number on aicost.ai is verified by 11 independent audit layers that run every day at 03:30 EDT — covering structural integrity, math correctness, source-side freshness, and cross-source agreement. We publish today's snapshot date and per-vendor verification timestamps below so you can verify any number yourself.

81
calculators
31
vendors verified
62
cited claims
715
days of history
11
audit layers
See the 8 audit layers
  • Layer 1: Architecture (12 structural invariants)
  • Layer 2: Smoke test (every calc page renders)
  • Layer 3: Golden values (math correctness vs reference)
  • Layer 4: Source resilience (independent reference data sources reachable)
  • Layer 5: Math gotchas (static code analysis)
  • Layer 6: Hybrid reconciliation (cross-source agreement)
  • Layer 7: Drift detection (day-over-day price changes)
  • Layer 8: Vendor cache (per-vendor freshness wiring)
  • Layer 9: Cross-vendor reachability (live vendor pricing page probes)
  • Layer 10: Rendered HTML drift (calc page DOM contracts, 45 pages daily)
  • Layer 11: Pricing freshness (cron heartbeat + per-vendor age tracking)

All 8 layers must pass before any pricing data is considered fresh. The infrastructure runs daily and publishes results to an internal dashboard. If any layer flags an issue, it is treated as stop-the-line work.

How this calculator sources its numbers

Every value falls into one of five categories. Numbers without an asterisk are vendor-published, directly observable, or computed by arithmetic on published data. Numbers marked with * are typical best-target values — we state the working range and invite you to override with your own number.

Vendor-published Directly from the vendor's pricing or docs page. No asterisk.
Published benchmark Independent benchmark (e.g. Chatbot Arena, ANN-Benchmarks, vLLM). Cited with date.
Research paper Peer-reviewed or widely-accepted research (e.g. LLMLingua, RAGAS).
Typical target * No single canonical source exists. We state the working range and explain why.
Computed Arithmetic on vendor-published values (e.g. batch discount × standard rate).

Any individual claim may also be tagged with * if its source has not yet been re-verified against the current vendor page — treat such claims as approximate until the next verification cycle resolves them.

Vendor verification freshness

Each vendor's pricing page is independently re-checked on a cadence ranging from daily to weekly. Below: when each relevant vendor was last verified by our automated pipeline.

anthropic
2026-10-26
verified today
by auto-pipeline
openai
2026-10-26
verified today
by auto-pipeline
xai
2026-10-26
verified today
by auto-pipeline
google-ai-plans-compare
2026-07-22
verified 96 days ago
by auto-grounded
azure-openai-pricing
2026-04-25
verified 184 days ago
anthropic-docs
2026-04-17
verified 192 days ago

Vendor-published values

Directly from the vendor's own docs. See per-vendor verification dates in the panel above.

Agentic workloads with stable system prompts and codebase context achieve 50ÔÇô80% cache hit rates.

Value
0.500000 fraction
Source
Anthropic ÔÇö Prompt Caching docs · Sun Mar 15 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Calculator default 0.50 is conservative. Production deployments with well-structured prompts often hit 0.70+. Vendor cache rates: Anthropic 10% of input, OpenAI 50%, Google 25% ÔÇö calculator uses published rates per vendor with NO 10% fallback.

Gemini Batch API provides a 50% cost reduction compared to realtime (Standard) pricing for input and output tokens.

“Batch API (50% cost reduction)”
Value
0.500000 fraction of base
Source
Google — Gemini Developer API pricing · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Discount applies to input and output token pricing across all models with Batch API support.

Prompt caching reduces costs by up to 90% for long prompts.

“reducing costs by up to 90% and latency by up to 85% for long prompts”
Value
0.900000 %
Source
Anthropic — Prompt Caching Launch · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to long prompts; specific use cases may vary.

Prompt caching reduces latency by up to 85% for long prompts.

“reducing costs by up to 90% and latency by up to 85% for long prompts”
Value
0.850000 %
Source
Anthropic — Prompt Caching Launch · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to long prompts; specific use cases may vary.

Anthropic offers a 1-hour cache duration at additional cost.

“If you find that 5 minutes is too short, Anthropic also offers a 1-hour cache duration at additional cost.”
Value
60.000000 minutes
Source
Anthropic — Prompt Caching documentation · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

The 1-hour cache duration is available on the Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, and Microsoft Foundry.

Cache read tokens are billed at 0.1 times the base input token price for most models.

“Cache reads are 0.1 times the base input tokens price (see the table footnote for per-model exceptions)”
Value
0.100000 fraction of base
Source
Anthropic — Prompt Caching documentation · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to all models except those explicitly listed with different multipliers (Claude Fable 5.1, Claude Mythos 5.1, Claude Opus 5.5, and Claude Sonnet 5.5).

Cache read tokens on Claude Fable 5.1 and Claude Mythos 5.1 are billed at 0.025 times the base input token price.

“Cache hits and refreshes on Claude Fable 5.1 and Claude Mythos 5.1 are priced at 0.025x the base input price.”
Value
0.025000 fraction of base
Source
Anthropic — Prompt Caching documentation · Wed Sep 30 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Specific exception for Claude Fable 5.1 and Claude Mythos 5.1 models.

Cache read tokens on Claude Opus 5.5 are billed at 0.05 times the base input token price.

“Cache hits and refreshes on Claude Opus 5.5 are priced at 0.05x the base input price.”
Value
0.050000 fraction of base
Source
Anthropic — Prompt Caching documentation · Wed Sep 30 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Specific exception for Claude Opus 5.5 model.

The default cache lifetime (TTL) is 5 minutes.

“By default, the cache has a 5-minute lifetime.”
Value
5.000000 minutes
Source
Anthropic — Prompt Caching documentation · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

The TTL is measured from the start of the request that writes or reads the cache entry.

1-hour cache write tokens are billed at 2 times the base input token price.

“1-hour cache write tokens are 2 times the base input tokens price”
Value
2.000000 multiplier
Source
Anthropic — Prompt Caching documentation · Wed Oct 07 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to all models supporting prompt caching.

1-hour cache write tokens are billed at 2 times the base input token price.

“1-hour cache write tokens are 2 times the base input tokens price”
Value
2.000000 multiplier
Source
Anthropic — Prompt Caching documentation · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to all models supporting prompt caching.

5-minute cache write tokens are billed at 1.25 times the base input token price.

“5-minute cache write tokens are 1.25 times the base input tokens price”
Value
1.250000 multiplier
Source
Anthropic — Prompt Caching documentation · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to all models supporting prompt caching.

Gemini prompt cache reads are billed at 50% of the base input token price for Standard, Batch, Flex, and Priority tiers.

“Batch API (50% cost reduction)”
Value
0.500000 fraction of base
Source
Google — Gemini Developer API pricing · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to all tiers with Batch API pricing; context caching is implicitly supported (explicit caching not supported).

Gemini context caching storage fee is $0.50 per 1,000,000 tokens per hour for Standard, Batch, Flex, and Priority tiers through December 31, 2026.

“$0.50 / 1,000,000 tokens per hour (storage price) through December 31, 2026.”
Value
0.500000 $/M tokens/hour
Source
Google — Gemini Developer API pricing · Wed Oct 21 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Storage fee applies to all tiers with context caching pricing; explicit caching not supported.

GPT-5.6 and later flagship models charge 0.1 times the uncached input token rate for cached reads.

“Subsequent reads cost 0.1× that rate on most of these models and 0.05× on GPT-6.1 Sol.”
Value
0.100000 fraction of base
Source
OpenAI — Prompt Caching documentation · Wed Oct 07 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

Applies to GPT-5.6 and later flagship models; cache reads are billed at 0.1× the uncached input token rate unless otherwise specified.

GPT-5.5 and GPT-5.5 Pro charge a model-dependent cached-input rate for reads.

“Cached-input rate”
Source
OpenAI — Prompt Caching documentation · Wed Oct 07 2026 00:00:00 GMT-0400 (Eastern Daylight Time)

The exact fraction is model-dependent and not explicitly stated in the documentation; refer to the model-specific pricing page for details.

Computed and derived values

Arithmetic on other verified numbers. Formulas shown.

Prompt caching saves money only when hit rate exceeds provider-specific break-even threshold*

Derived: Anthropic break-even = 0.25 / (1.0 - 0.10 + 0.25) = 21.7%. OpenAI/Google = 0% (no write premium). Assumes steady-state usage ? one-shot workloads may never amortize cache writes.

Method reference: Anthropic ? Prompt caching launch

Vendor pricing pages referenced

All vendor-published prices used by this calculator are sourced from the pages below. See the verification panel above for when each was last re-checked.

See an error or stale value?

We treat methodology as a living document. If a price is wrong, a benchmark is outdated, or you have a better citation, let us know and we will verify within 48 hours.

Email [email protected] →
AICost.ai pricing data keeps our calculations accurate across top AI and cloud vendors · last updated 2026-10-26

Methodology

  • All prices are USD per 1 million tokens, current as of 2026-10-26.
  • Vendor-published values have no mark. Inferred/extrapolated values are marked with * and listed below.
  • Batch API discounts are 50% off standard rates across providers that offer Batch mode.
  • Prompt caching discounts vary by provider (typically 80-90% off cached input tokens).
  • Regional data-residency surcharges (Anthropic 1.1x, OpenAI 1.1x, Google regional tiers) are NOT included in base rates.
  • Long-context pricing tiers apply when input exceeds model threshold.
  • Embedding prices are input-only (no output tokens generated).

Primary sources

Last-verified date is the most recent successful daily snapshot (aicost_pricing_snapshots) or, when no snapshot exists yet, the latest successful crawler run (aicost_crawler_runs). 10 of 10 vendors are currently verified. Aggregator services (TokenCost, AI Pricing Guru, etc.) are not listed.

Anthropic
2026-10-26
https://www.anthropic.com/pricing
Daily snapshot since Sep 2023 · 712 days captured
Anthropic Docs
2026-10-26
https://platform.claude.com/docs/en/about-claude/pricing
Daily snapshot since Sep 2023 · 712 days captured
OpenAI
2026-10-26
https://openai.com/api/pricing/
Daily snapshot since Sep 2023 · 713 days captured
Google AI
2026-10-26
https://ai.google.dev/gemini-api/docs/pricing
Daily snapshot since Dec 2023 · 688 days captured
Google Vertex
2026-10-26
https://cloud.google.com/vertex-ai/generative-ai/pricing
Daily snapshot since Dec 2023 · 688 days captured
DeepSeek
2026-10-26
https://api-docs.deepseek.com/quick_start/pricing
Daily snapshot since May 2024 · 627 days captured
xAI
2026-10-26
https://x.ai/api
Daily snapshot since Nov 2024 · 545 days captured
Mistral
2026-10-26
https://mistral.ai/pricing
Daily snapshot since Dec 2023 · 686 days captured
Cohere
2026-10-26
https://cohere.com/pricing
Daily snapshot since Sep 2023 · 712 days captured
Voyage AI
2026-10-26
https://docs.voyageai.com/docs/pricing

Inferred values (marked with * in calculator tables)

Derived from industry conventions, not directly published by the vendor. Typical conventions: cached input = 10% of base (90% off), Batch API = 50% of base (50% off).

Vendor / Model Field Why it’s inferred
Anthropic — Claude Sonnet 4.6 cachedInput Derived at 10% of input rate — Anthropic publishes 90% cache-hit discount on this tier.
Anthropic — Claude Sonnet 4.5 cachedInput Derived at 10% of input rate; same 90% cache-hit convention as Sonnet 4.6.
Anthropic — Claude Sonnet 4.5 batchInput Derived at 50% of standard input — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Sonnet 4.5 batchOutput Derived at 50% of standard output — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Haiku 4.5 cachedInput Derived at 10% of input rate — Anthropic 90% cache-hit discount convention.
OpenAI — GPT-5.4 Mini cachedInput Derived at 10% of input — OpenAI documents automatic 90% discount on cache hits across GPT-5.x tier.
OpenAI — GPT-5.4 Nano cachedInput Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Nano batchInput Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Nano batchOutput Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro cachedInput Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Pro batchInput Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro batchOutput Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.2 cachedInput Derived at 10% of input; no residency uplift.
OpenAI — GPT-5.2 batchInput Derived at 50% of input.
OpenAI — GPT-5.2 batchOutput Derived at 50% of output.
OpenAI — GPT-5 cachedInput Derived at 10% of input.
OpenAI — GPT-5 batchInput Derived at 50% of input.
OpenAI — GPT-5 batchOutput Derived at 50% of output.
OpenAI — GPT-5.5 Pro cachedInput Derived at 10% of input — OpenAI does not publish a cached rate for *-pro models; using the family convention.
OpenAI — GPT-5.5 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5.5 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5.2 Pro cachedInput Derived at 10% of input — pro-tier convention.
OpenAI — GPT-5.2 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5.2 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5.1 batchInput Derived at 50% of input.
OpenAI — GPT-5.1 batchOutput Derived at 50% of output.
OpenAI — GPT-5 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5 Nano cachedInput Derived at 10% of input.
OpenAI — GPT-5 Nano batchInput Derived at 50% of input.
OpenAI — GPT-5 Nano batchOutput Derived at 50% of output.
Google — Gemini 3 Flash cachedInput Derived at 10% of input — Google caching discount convention ~90%.
Google — Gemini 3.1 Flash-Lite cachedInput Derived at 10% of input — Google caching convention.
Google — Gemini 3.1 Flash-Lite batchInput Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 3.1 Flash-Lite batchOutput Derived at 50% of output — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Pro cachedInput Derived at 10% of input.
Google — Gemini 2.5 Flash cachedInput Derived at 10% of input.
Google — Gemini 2.5 Flash-Lite cachedInput Derived at 10% of input — Google caching convention.
Google — Gemini 2.5 Flash-Lite batchInput Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Flash-Lite batchOutput Derived at 50% of output — Google Batch API uniform 50% discount.
xAI — Grok 4 (legacy) cachedInput Extrapolated at 25% of base.

Pricing is cross-verified against the LiteLLM community registry when available. Daily snapshots are kept in aicost_pricing_snapshots; every change is logged to aicost_price_changelog with old & new values for full audit trail. Read the full methodology →