{
  "schemaVersion": 1,
  "id": "provider-cache-hit-rate/2026-05",
  "series": "provider-cache-hit-rate",
  "title": "Prompt cache hit rate by provider, May 2026",
  "topic": "caching",
  "question": "What share of input tokens do LLM providers serve from prompt cache?",
  "questions": [
    "Which LLM provider has the highest prompt cache hit rate?",
    "What is a typical prompt caching hit rate in production?",
    "How much of my input is served from cache on Anthropic, OpenAI or Bedrock?"
  ],
  "description": "Share of input tokens served from the provider's prompt cache, per provider and month.",
  "period": "2026-05",
  "period_start": "2026-05-01",
  "period_end": "2026-05-31",
  "updated": "2026-11-11",
  "is_latest": false,
  "headline": "DeepSeek served 83.8% of input tokens from cache in May 2026, the highest among major providers. The median major provider served 40.9%.",
  "key_findings": [
    "DeepSeek leads at 83.8%; Novita is lowest at 5.3%.",
    "2 of 10 major providers served at least 80% of input tokens from cache."
  ],
  "columns": [
    {
      "key": "provider",
      "label": "Provider",
      "unit": "text"
    },
    {
      "key": "cache_hit_rate",
      "label": "Cache hit rate",
      "unit": "fraction",
      "description": "Cached input tokens divided by input tokens."
    },
    {
      "key": "request_share",
      "label": "Traffic share",
      "unit": "fraction"
    },
    {
      "key": "sample_band",
      "label": "Organizations (band)",
      "unit": "text",
      "description": "Number of distinct organizations behind the row, in bands."
    }
  ],
  "rows": [
    {
      "provider": "deepseek",
      "cache_hit_rate": 0.8382,
      "request_share": 0.0394,
      "sample_band": "100-499",
      "provider_label": "DeepSeek"
    },
    {
      "provider": "anthropic",
      "cache_hit_rate": 0.806,
      "request_share": 0.2447,
      "sample_band": "500+",
      "provider_label": "Anthropic"
    },
    {
      "provider": "azure",
      "cache_hit_rate": 0.6772,
      "request_share": 0.036,
      "sample_band": "100-499",
      "provider_label": "Azure"
    },
    {
      "provider": "vertex_claude",
      "cache_hit_rate": 0.4909,
      "request_share": 0.011,
      "sample_band": "100-499",
      "provider_label": "Vertex (Claude)"
    },
    {
      "provider": "bedrock",
      "cache_hit_rate": 0.4572,
      "request_share": 0.0481,
      "sample_band": "100-499",
      "provider_label": "Bedrock"
    },
    {
      "provider": "openai",
      "cache_hit_rate": 0.3601,
      "request_share": 0.3134,
      "sample_band": "500+",
      "provider_label": "OpenAI"
    },
    {
      "provider": "xai",
      "cache_hit_rate": 0.3239,
      "request_share": 0.0198,
      "sample_band": "100-499",
      "provider_label": "xAI"
    },
    {
      "provider": "vertex_gemini",
      "cache_hit_rate": 0.1942,
      "request_share": 0.221,
      "sample_band": "500+",
      "provider_label": "Vertex (Gemini)"
    },
    {
      "provider": "deepinfra",
      "cache_hit_rate": 0.1263,
      "request_share": 0.0054,
      "sample_band": "25-99",
      "provider_label": "DeepInfra"
    },
    {
      "provider": "novita",
      "cache_hit_rate": 0.0526,
      "request_share": 0.0406,
      "sample_band": "100-499",
      "provider_label": "Novita"
    }
  ],
  "totals": {
    "withheld_request_share": 0.0301
  },
  "floors": {
    "min_requests": 20,
    "min_orgs": 5,
    "max_top_org_share": 0.8
  },
  "methodology": [
    "Cache hit rate is cached input tokens divided by input tokens, the same definition /rankings uses.",
    "Providers that report no cached tokens are left out of the chart. Only providers with at least 0.5% of published traffic are shown.",
    "A row is published only when it has at least 20 requests, at least 5 distinct organizations, and no single organization above 80% of its traffic. These are the same floors /rankings uses. Rows below the floors are withheld, not estimated.",
    "Only shares, rates, medians and organization bands are published. Absolute request, token and spend totals stay private."
  ],
  "definitions": [
    {
      "term": "Cache hit rate",
      "text": "Cached input tokens / input tokens, summed over the month."
    }
  ],
  "caveats": [
    "Hit rate depends on workload. Long agent sessions with a stable prefix cache far better than one-off requests.",
    "Providers report cached tokens differently. Some only count explicit cache reads."
  ],
  "source": "Requesty production gateway",
  "license": {
    "name": "CC BY 4.0",
    "url": "https://creativecommons.org/licenses/by/4.0/"
  },
  "citation": {
    "apa": "Requesty (2026). Prompt cache hit rate by provider, May 2026. Requesty Data. https://www.requesty.ai/data/provider-cache-hit-rate/2026-05",
    "bibtex": "@misc{requesty_provider_cache_hit_rate_2026_05,\n  author       = {{Requesty}},\n  title        = {Prompt cache hit rate by provider, May 2026},\n  year         = {2026},\n  howpublished = {\\url{https://www.requesty.ai/data/provider-cache-hit-rate/2026-05}},\n  note         = {Requesty Data}\n}"
  },
  "permalink": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-05",
  "latest_url": "https://www.requesty.ai/data/provider-cache-hit-rate",
  "previous": null,
  "next": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-06",
  "periods": [
    {
      "period": "2026-05",
      "url": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-05",
      "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-05/data.json"
    },
    {
      "period": "2026-06",
      "url": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-06",
      "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-06/data.json"
    },
    {
      "period": "2026-07",
      "url": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-07",
      "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-07/data.json"
    },
    {
      "period": "2026-08",
      "url": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-08",
      "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-08/data.json"
    },
    {
      "period": "2026-09",
      "url": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-09",
      "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-09/data.json"
    },
    {
      "period": "2026-10",
      "url": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-10",
      "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-10/data.json"
    }
  ],
  "downloads": {
    "json": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-05/data.json",
    "csv": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-05/data.csv",
    "markdown": "https://www.requesty.ai/data/provider-cache-hit-rate/2026-05/data.md"
  },
  "related": [
    "https://www.requesty.ai/data/provider-error-rates",
    "https://www.requesty.ai/data/provider-latency"
  ],
  "rankings": "https://www.requesty.ai/rankings"
}