ByteCosts
how-llm-providers-price-prompt-caching.mdx

AI Economics

How the Major LLM Providers Price Prompt Caching

Last updated 2026-07-19 · ByteCosts

Cache reads can be much cheaper than uncached input, but prompt caching does not have one universal break-even point. The result depends on the write premium, repeated prefix tokens, measured hit rate, TTL, and provider eligibility rules, plus any storage charge. Compare the full cached path against ordinary input for the same workload.

Quick answer

In the named model snapshots below, Anthropic, OpenAI, and Google price a cache read at about a tenth of fresh text input. DeepSeek prices a V4-Flash cache hit at roughly a fiftieth of its cache miss. Those read discounts are not the whole bill: Anthropic charges a write premium, Google adds context-cache storage, OpenAI requires eligible exact-prefix reuse, and DeepSeek’s blended rate depends on observed hits and misses. If you need the mechanism before the commercial comparison, start with what prompt caching is.

The pricing pages, side by side

ProviderModel or planUncached input priceCached input or cache-read priceWrite / storage feeEligibility / minimum prompt lengthCache mechanicsBest-fit workloadOfficial sourceLast checked
AnthropicClaude Sonnet 4.6$3.00/MTok$0.30/MTok cache read5-minute cache write $3.75/MTok; 1-hour write $6.00/MTokModel-specific minimums and explicit cache breakpointsCache reads are 0.1x base input; writes cost more and expire by TTLReused system prompts, tools, examples, and long reference contextAnthropic pricing and prompt caching2026-06-15
OpenAIgpt-5.5$5.00/MTok$0.50/MTok cached inputno separate cache-write fee listedAutomatic for prompts of at least 1,024 tokensRequires exact prefix matches; static content should come first; cached prompts still count toward rate limits and do not change output tokensLong, repeated prefixes with static content first and variable content lastOpenAI pricing and prompt caching2026-06-15
GoogleGemini 2.5 Flash, Standard$0.30/MTok text/image/video$0.03/MTok context caching$1.00 per 1M tokens per hour of context-cache storageContext cache must be created and kept warmStorage time is billed separately, so slow reuse can reduce savingsLarge shared context reused across many calls inside the cache windowGemini API pricing2026-06-15
DeepSeekDeepSeek-V4-Flash$0.14/MTok cache miss$0.0028/MTok cache hitno separate write/storage fee listed on the pricing tableProvider bills hit and miss input separatelyBlended cost depends on observed cache-hit rate; deepseek-chat and deepseek-reasoner are compatibility aliases until deprecationVery repetitive prefixes where cache hits are commonDeepSeek pricing2026-06-15

Worked break-even example

This example uses only the Anthropic Claude Sonnet 4.6 rates in the table above. It illustrates the arithmetic; it is not a universal savings forecast.

Assumptions: The workload sends the same eligible 1 million-token prefix twice inside the 5-minute TTL. The first request writes the prefix at $3.75/MTok and the second gets a full cache read at $0.30/MTok. There are no expiry, eviction, or eligibility failures, so the repeated-prefix hit rate after the write is 100%. Variable input and output tokens are excluded because they are identical on both sides of this comparison.

  • Without caching: 2 × $3.00 = $6.00 for the repeated prefix.
  • With caching: $3.75 + $0.30 = $4.05 for one write and one read.
  • Break-even check: At one request, the $3.75 write versus $3.00 ordinary input does not break even. At two requests, this modeled cache path passes the break-even threshold and moves below the $6.00 uncached path. Algebraically, $3.75 + (N - 1) × $0.30 < N × $3.00, which requires N > 1.28; the first whole-request count that clears the threshold is two.

Your result can reverse with a lower hit rate, a prefix change, or expiry before reuse. Put your own token volume, hit rate, and request count into the prompt-cache savings calculator, then compare complete workload variants in Scenario Studio.

Where they mostly agree

Three of the four land cache reads at about a tenth of fresh text input in the example rows above. Anthropic states it as a multiplier (0.1x), while Google ($0.03 against $0.30) and OpenAI ($0.50 against $5.00) state it as a discounted per-token rate. DeepSeek goes further still, pricing a V4-Flash cache hit at roughly a fiftieth of a cache miss. Either way, if your workload re-reads a large, stable context, the blended cache-read line can dominate the input economics.

Where they differ

  • Anthropic charges to write. Reads are cheap at 0.1x, but writing tokens into the cache costs 1.25x base input for the 5-minute window and 2x for the 1-hour window. You pay up front, so caching pays off only when the same context is reused enough times to amortize the write.
  • Google charges rent. Cached input is cheap at $0.03, but Google bills $1.00 per 1M tokens per hour to keep the cache warm. A large context held open across a slow session can cost more in storage than it saves on reads.
  • OpenAI charges nothing extra, but eligibility is narrow. Its docs say caching works automatically and has no additional fees, but only eligible long prompts are cached. Caching is available for prompts of 1,024 tokens or more, cache hits require exact prefix matches, cached prompts still count toward TPM rate limits, and caching does not change output-token generation. Static content should sit at the beginning of the prompt and variable content at the end.
  • DeepSeek prices the gap. The cache hit at $0.0028 against a $0.14 miss is the steepest spread of the four, rewarding reuse heavily and punishing a cold cache.

How to read this for your own bill

The right comparison is never the input sticker price; it is the blended rate your workload actually hits once cache writes, reads, and any storage are counted at your reuse ratio. Pull the current per-token rates from the ByteCosts AI Provider Pricing Index and model your own cache-hit ratio in the AI cost calculator before you pick a provider on its headline number.

Key takeaways

  • For repeated long-prefix workloads, Anthropic, OpenAI, and Google price cache reads at about a tenth of fresh text input in the examples above; DeepSeek goes further, to roughly a fiftieth (cache hit versus miss).
  • They recover the cost differently: Anthropic adds a write premium (1.25x), Google adds storage ($1.00 per 1M tokens per hour), OpenAI adds nothing, DeepSeek prices a steep hit-miss gap ($0.0028 vs $0.14).
  • Compare on the blended rate at your reuse ratio, token length, TTL, and provider eligibility rules, not the input sticker price.

Pricing snapshot limits

The prices in this article are committed planning snapshots, not live provider data. Provider pricing, plan limits, regions, cache rules, model aliases, discounts, taxes, and marketplace billing can change after the last-checked date. Where this article shows a comparison, the comparison is scoped to the named model rows and provider pages. It is not a billing guarantee.

Evidence and last-checked dates

Claim areaSource typeSourceLast checkedNotes
Anthropic cache read/write multipliersOfficial provider pricing and caching docsClaude API pricing and prompt caching2026-06-15Write/read prices vary with model base input price and TTL.
OpenAI cached input and mechanicsOfficial provider pricing and caching docsOpenAI pricing and prompt caching2026-06-15Caching requires eligible long prompts and exact prefix reuse; no separate write fee found.
Google context cachingOfficial provider pricingGemini API pricing2026-06-15Gemini 2.5 Flash Standard text/image/video row; storage rent is separate.
DeepSeek cache hit/miss pricingOfficial provider pricingDeepSeek pricing2026-06-15V4 Flash and V4 Pro rows expose cache-hit and cache-miss prices directly.

Sources

How the Major LLM Providers Price Prompt Caching. ByteCosts. Updated 2026-07-19. https://bytecosts.com/blog/how-llm-providers-price-prompt-caching/

Sources