{
  "guide": "batch-api-economics",
  "title": "Batch API economics: when the -50% applies",
  "url": "https://optimtoken.optimnow.io/guides/batch-api-economics",
  "page": "https://optimtoken.optimnow.io/guides/batch-api-economics",
  "dataSource": "llm",
  "snapshotDate": "2026-11-12",
  "dateModified": "2026-11-12",
  "note": "Figures recomputed from the OptimToken catalogue at each build, the same ones the page prints. A figure quoted elsewhere is as of the day it was read; this file carries the current one.",
  "figures": {
    "unit": "batch discount as a fraction of list price on a 30/70 input-output blend; cost in dollars per request",
    "models": 311,
    "publishingBatch": 74,
    "rated": 74,
    "medianBatchDiscount": 0.5,
    "bands": {
      "Exactly 50% off": 63,
      "More than 50% off": 2,
      "Partial discount": 9,
      "No discount": 0,
      "Batch priced above list": 0
    },
    "exceptions": [
      {
        "model": "gpt-oss-120b",
        "provider": "OpenAI",
        "discount": 0.7762
      },
      {
        "model": "DeepSeek V4.1 Flash 1017",
        "provider": "DeepSeek",
        "discount": 0.6918
      },
      {
        "model": "Qwen3.8 Max (0902)",
        "provider": "Alibaba",
        "discount": 0.425
      },
      {
        "model": "Qwen3.7 Plus",
        "provider": "Alibaba",
        "discount": 0.2812
      },
      {
        "model": "MiniMax M3",
        "provider": "MiniMax",
        "discount": 0.2654
      },
      {
        "model": "Inkling Small",
        "provider": "thinkingmachines",
        "discount": 0.2
      },
      {
        "model": "Grok 4.3",
        "provider": "xAI",
        "discount": 0.2
      },
      {
        "model": "Muse Glimmer 30B",
        "provider": "meta",
        "discount": 0.2
      },
      {
        "model": "Grok 4.6",
        "provider": "xAI",
        "discount": 0.2
      },
      {
        "model": "Grok 4.7",
        "provider": "xAI",
        "discount": 0.2
      },
      {
        "model": "gpt-oss-20b",
        "provider": "OpenAI",
        "discount": 0.144
      }
    ],
    "coverageByMaker": [
      {
        "maker": "OpenAI",
        "models": 49,
        "withBatch": 37
      },
      {
        "maker": "Alibaba",
        "models": 40,
        "withBatch": 2
      },
      {
        "maker": "Google",
        "models": 28,
        "withBatch": 10
      },
      {
        "maker": "Mistral",
        "models": 20,
        "withBatch": 1
      },
      {
        "maker": "Meta",
        "models": 17,
        "withBatch": 1
      },
      {
        "maker": "Zhipu",
        "models": 17,
        "withBatch": 0
      },
      {
        "maker": "Anthropic",
        "models": 16,
        "withBatch": 16
      },
      {
        "maker": "DeepSeek",
        "models": 14,
        "withBatch": 1
      },
      {
        "maker": "Moonshot",
        "models": 8,
        "withBatch": 0
      },
      {
        "maker": "xAI",
        "models": 7,
        "withBatch": 3
      }
    ],
    "workedExample": {
      "profile": "invoiceProcessing",
      "inputTokens": 1500,
      "outputTokens": 600,
      "requestsAMonth": 100000,
      "models": [
        {
          "model": "Claude Sonnet 5",
          "provider": "Anthropic",
          "listPerRequest": 0.009,
          "batchPerRequest": 0.0045
        },
        {
          "model": "GPT-5.5",
          "provider": "OpenAI",
          "listPerRequest": 0.0255,
          "batchPerRequest": 0.0127
        },
        {
          "model": "GLM 5.3 Flash",
          "provider": "Zhipu",
          "listPerRequest": 0.0001,
          "batchPerRequest": null
        }
      ],
      "counterExample": {
        "model": "GLM 5.3 Flash",
        "provider": "Zhipu",
        "rule": "cheapest model with no published batch rate that clears the preset's quality bar and undercuts the first model's batch cost at list price; deprecated models and outlier listings left out",
        "qualityBar": {
          "label": "Standard",
          "score": 1420
        },
        "eloScore": 1471,
        "listPerRequest": 0.0001,
        "againstBatchPerRequest": 0.0045
      }
    }
  }
}