Is Claude or OpenAI prompt caching cheaper in practice?

Consensus at this check

Short answer

Neither is universally cheaper. In practice:

Claude is often cheaper for dense, high-hit-rate workloads, especially repeated large prompts within its 5-minute cache window.
OpenAI can be cheaper for sporadic traffic, particularly when cache misses do not incur a write premium, or when the chosen OpenAI model has a lower base input price.
The biggest factors are the specific models, cache-hit rate, request timing, and input/output-token mix—not just the advertised cache discount.

Pricing mechanics

Claude generally charges:

Normal input price for uncached tokens
1.25× normal input price for a 5-minute cache write
2× for a 1-hour cache write
About 0.1× normal input price for cache reads.anthropic

OpenAI’s cached-input pricing is model-dependent. Many models price cached input at approximately 0.1× standard input, while some tiers are closer to 0.5×; cache-write treatment also varies by model and API configuration.openai OpenAI caching is generally automatic, based on matching prompt prefixes, with retention depending on the model/API; current documentation describes options around a 30-minute TTL.openai

Normalized example

Let the uncached cost of a reusable prompt prefix be \(B\).

For Claude’s 5-minute cache, with \(R\) subsequent cache hits:

\[ \text{cost} \approx 1.25B + 0.10BR \]

Without caching:

\[ \text{cost} = B(1+R) \]

After one cache hit, Claude costs \(1.35B\) instead of \(2B\), saving about 32.5%. With many hits, the savings approach 90%, after accounting for the initial write premium.

If OpenAI charges \(1.0B\) on a miss and \(0.1B\) per cached read, it is slightly cheaper than Claude on each cache cycle. If it charges \(0.5B\) per cached read, Claude becomes substantially cheaper after repeated hits. Thus, the conclusion changes with the OpenAI model’s cached-input rate.

When Claude tends to win

Claude is usually attractive when:

Requests arrive frequently within five minutes.
The reusable prefix is large.
You have a high cache-hit rate.
You can explicitly structure stable system instructions, tools, documents, or context for caching.
You are comparing a lower-priced Claude model with a more expensive OpenAI model.

Claude’s explicit cache breakpoints can also be useful for layered prompts containing stable tools, documents, and changing conversation state.

When OpenAI tends to win

OpenAI may be cheaper or safer when:

Requests are separated by more than Claude’s 5-minute default window.
Cache misses are common.
The selected OpenAI model has a lower base input price.
Its model/API configuration does not charge a cache-write premium.
You benefit from automatic prefix caching without having to manage cache breakpoints.openai

Practical rule

For a repeated 100,000-token prefix:

High-frequency agent loop: Claude often has the lower total input cost, provided its cache remains warm.
Occasional requests: OpenAI may win because Claude’s cache-write premium can be paid repeatedly without enough reads to amortize it.
Model-to-model comparison: Compare actual per-token prices first. A 90% discount on an expensive model can still cost more than a 90% discount on a cheaper one.

Finally, benchmark cost per completed task, including output tokens, reasoning tokens, tool calls, cache writes, and cache reads. Prompt caching only discounts repeated input; it does not necessarily dominate the total bill.

Sources

Position Map

Where the models stand

Each row is one part of the answer. The cards show the distinct positions; the model chips show who supports each one.

0/100 Direction Shift · Stable

Models disagree

OpenAI provides a 50% discount on cached reads without write fees versus matching Claude's 90% read discount and 1.25x write fee.

Position 1

OpenAI provides a 50% discount and charges no write surcharge.

Gemini

Position 2

Older OpenAI models offer 50% off with no write fee, whereas newer models offer 90% off and some charge a 1.25x write fee.

DeepSeek

Position 3

OpenAI's current models match Claude with a 0.1x read price and 1.25x write surcharge.

OpenAI

See how each model moved across checks

Model position movement by watch date
Model	Aug 11	Aug 18	Aug 25	Sep 01
OpenAI
Grok			—	—
Gemini	—
DeepSeek	—	—	—

Cite this answer

consens.io. (2026-08-25). Consensus answer to "Is Claude or OpenAI prompt caching cheaper in practice?". Models consulted: OpenAI: gpt-5.6-luna, Google Gemini: gemini-3.7-flash. Consensus model: gpt-5.6-luna. Sources: https://docs.anthropic.com/en/docs/about-claude/pricing?4810b549_page=3&73cdfb14_page=2&939688b5_page=1&e768fcd2_page=2&utm_source=openai, https://platform.openai.com/pricing?model=contentfilter-alpha-001&utm_source=openai, https://developers.openai.com/api/reference/cli/resources/responses/methods/create?utm_source=openai, https://openai.com/index/api-prompt-caching/?utm_source=openai, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFgnrhvywIqKqMtiigG73IQQUCB43NvWWMeAVWM8IPKRR7MsaaduIq7as1KZ2su0Q01xQPyynl17MhxVc90vxMB4aw_66dWZxHsSErqMfluN848YQmsVOhWC9JHrq2PCdXOkNZUtg==, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEXwJfqP1qchYfAn35RoVPhtlKasaHgqP7fhjkqnI_UJKWozE3jPZ8FMhxDpEKM1kl1Q2NMcYPkrdOmPZRh9SP1MNsjcD3NV6iYX3tngOugM0UE8FxLwuMqFMqqeMleHu3qHHiVytHvOIWQMo4jDg==, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEEqpfxyshGDLW0a6gBU7zJCw9NvE4nuSteyDKk47pYBzDk8r-Yph7Ae9pMV_XAcQhljQUcbe6eH8-lkxW4vpd01V7gfCx8eLM5O6pc686drjx1F_u-f2ryb35TY6PA9ukKdN3bY__fNtA4uv-ihbOZ, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFhft-bgIL6vy4HlsGq9nPYIUP-aRqyzyPczCqMqSRHYyavoHH9rtkqDqjAKCoaV8pdz3v7C4m4hoJwV91uUhuKpMrN2VOjT-ADfWaHPJ7YviPciVBr6IQBZRc-G27emzEBAEbPfaPznlrWgDPnsCqWXW-ZRNkh0H1FG9rFCSqZ1hyhk7bGcvohC2AMgSqgBSnQfH3iO7c=, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHTPQD3TBgEPOJfM8xiCXTH2Z6_8Xmpi1fBDy5IdAzrIol4n6hrHqhArcipOKi65GLjo-MLWRj8UUtZpARuwQGuwkIUGHtxZm5k3_HUG2R5wDGe9sWgNpp8zJbzwJtRQ15teL6je8n73MAkxC5GYhfDVv4G4xP1kQfoJWCbRcPSWAK0Xp0iePwG1j5zeajeVZLhi8MdvYP4maDbZg==, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFmNdaTF24JZqTPAn3XfBIrAJm77HLYSGgJ0nQaUR8CtMsVlbY172Gtqqg8BOMS-DVjvh4KpZXY4x-rb_tlbR2nASk3BTIViCibfPxAlTpHd5JzqHkHeehPhngEqk_nHZsiqp8GLZ-BtBGxbXdO4Q8BLGzbjbXuMHRbT1vZoT25n74VtFyRqYaa, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEWUMYeK2xUIZcXaEGVgJdmuoGWxAeQc1cy7vexx58ssyjquczer0791UlfREtcc_dwPGrKfQ5qcFHzdRxp1xZzj62vK82i5jUbLqqkMe6PIcqZ0ZVRgBcXxRpAosgCZE2QR33OUsClo0ny6rNmWY350krqxYyft9IkrvAkLA==, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQFlYRStIOfYNeNU4ME134h3Gdpad9peev1fh0SA5y_xQJpxNixN8q3N1kKVoktOYDa7lpZnDUfl29jwd3SyX2Rxpy_-GNXUj2oX1fbReX4gLwVVdX4vME4WwO2x8JoKANTArhsv6esq9Wjs, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGbkOnGCm7h2hYyLWrGR_fjUL8EQzuZpO8zO1RsJ15BuBhvA8wnfb2ggtzu_QHKFT_wEek-wIqJofROkn9M5Nj-2Dix26fuC5g1V5Jd5XFEjS_B6EPJ6udXtcHEfuIYclEy9cU5UfzxeAD-IRzkCvi-SgrGQd7z9vlMoQ==, https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEjLC0sjWlDrdwNZiEzTsQAg9urzu4-Pgr4TO-wtY41XpSTThKzi4FLoLxMwYjKGoX9CdV9TFVGkJ9u2GVqESAnwn23ERe0RITR3IZAPuxRTaGMX8hhn-IUtLTIShh5VCJDng== Retrieved from https://www.consens.io/s/is-claude-or-openai-prompt-caching-cheaper-in-practice-KaTChmKxXX90G0vc?version=6d4cf47b15aba9d150c496f6

Ask your own question

Consensus Watch

Run history

64/100 latest agreement

View the full agreement chart

Agreement over time

How strongly the models support the same claims. Every point links to its run below.

Checks

Newest first. Open any saved result to read the full consensus from that date.

2026-09-01 Stable
64/100 agreement
No meaningful movement detected in this check.
Open this consensus
2026-08-25 Stable
75/100 agreement
Restated, not moved: The consensus shifted from stating OpenAI is usually cheaper in default practice to clarifying that neither is universally cheaper. It adds explicit mathematical breakdowns, details on OpenAI's tiered cache read discounts (0.1x to 0.5x) vs. Claude's write premiums/5-minute window, and broader workload nuances.
Open this consensus
2026-08-18 Meaningful change
90/100 agreement
Refined the pricing comparison to clarify that both providers offer similar cache read discounts (~90%) and that total cost differences often stem from underlying base model prices, while maintaining the same core conclusions regarding which provider to choose for different workload patterns.
Open this consensus
2026-08-11 Stable
64/100 agreement
No meaningful movement detected in this check.
Open this consensus