{
  "title": "Brass AI price index",
  "license": "CC BY 4.0, credit \"Brass AI price index, brass.adorellc.pro\"",
  "unit": "USD per million tokens, list price, three input tokens per output token",
  "as_of": "2026-10-01",
  "method": "Cheapest published list price for a model scoring at least as well as the original GPT-4 on GPQA Diamond (scores from Epoch AI), blended as three input tokens per output token.",
  "rows": [
    {
      "series": "GPT-4-level work",
      "date": "2023-03-14",
      "model": "GPT-4 (gpt-4-0314)",
      "input": 30,
      "output": 60,
      "blended": 37.5,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20230315033154/https://openai.com/pricing",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2023-11-06",
      "model": "GPT-4 Turbo (gpt-4-1106-preview)",
      "input": 10,
      "output": 30,
      "blended": 15,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20231106184005/https://openai.com/blog/new-models-and-developer-products-announced-at-devday",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2024-02-26",
      "model": "Mistral Large (mistral-large-2402)",
      "input": 8,
      "output": 24,
      "blended": 12,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20240226193838/https://docs.mistral.ai/platform/pricing/",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2024-03-04",
      "model": "Claude 3 Sonnet",
      "input": 3,
      "output": 15,
      "blended": 6,
      "status": "verified",
      "sources": [
        "https://www.anthropic.com/news/claude-3-family",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2024-03-13",
      "model": "Claude 3 Haiku",
      "input": 0.25,
      "output": 1.25,
      "blended": 0.5,
      "status": "verified",
      "sources": [
        "https://www.anthropic.com/news/claude-3-haiku",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2024-07-18",
      "model": "GPT-4o mini",
      "input": 0.15,
      "output": 0.6,
      "blended": 0.2625,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20240718174226/https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2024-08-12",
      "model": "Gemini 1.5 Flash (price cut)",
      "input": 0.075,
      "output": 0.3,
      "blended": 0.13125,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20240815204013/https://developers.googleblog.com/en/gemini-15-flash-updates-google-ai-studio-gemini-api/",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2025-01-30",
      "model": "Qwen Turbo (qwen-turbo-2024-11-01)",
      "input": 0.05,
      "output": 0.2,
      "blended": 0.0875,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20250130085941/https://www.alibabacloud.com/help/en/model-studio/getting-started/models",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "GPT-4-level work",
      "date": "2026-07-27",
      "model": "Qwen3.7 Flash",
      "input": 0.03,
      "output": 0.13,
      "blended": 0.055,
      "status": "verified",
      "sources": [
        "https://www.alibabacloud.com/help/en/model-studio/model-pricing",
        "https://epoch.ai/data/benchmark_data.zip"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2023-03-14",
      "model": "GPT-4",
      "input": 30,
      "output": 60,
      "blended": 37.5,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20230315033154/https://openai.com/pricing"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2023-11-06",
      "model": "GPT-4 Turbo",
      "input": 10,
      "output": 30,
      "blended": 15,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20231106184005/https://openai.com/blog/new-models-and-developer-products-announced-at-devday"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2024-05-13",
      "model": "GPT-4o",
      "input": 5,
      "output": 15,
      "blended": 7.5,
      "status": "verified",
      "sources": [
        "https://web.archive.org/web/20240513221325/https://openai.com/api/pricing/"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2025-02-27",
      "model": "GPT-4.5 preview",
      "input": 75,
      "output": 150,
      "blended": 93.75,
      "status": "verified",
      "sources": [
        "https://developers.openai.com/api/docs/models/gpt-4.5-preview"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2025-08-07",
      "model": "GPT-5",
      "input": 1.25,
      "output": 10,
      "blended": 3.4375,
      "status": "verified",
      "sources": [
        "https://developers.openai.com/api/docs/models/gpt-5"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2025-12-11",
      "model": "GPT-5.2",
      "input": 1.75,
      "output": 14,
      "blended": 4.8125,
      "status": "verified",
      "sources": [
        "https://developers.openai.com/api/docs/models/gpt-5.2"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2026-03-05",
      "model": "GPT-5.4",
      "input": 2.5,
      "output": 15,
      "blended": 5.625,
      "status": "verified",
      "sources": [
        "https://developers.openai.com/api/docs/models/gpt-5.4"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2026-04-24",
      "model": "GPT-5.5",
      "input": 5,
      "output": 30,
      "blended": 11.25,
      "status": "verified",
      "sources": [
        "https://developers.openai.com/api/docs/models/gpt-5.5"
      ]
    },
    {
      "series": "OpenAI flagship",
      "date": "2026-09-03",
      "model": "GPT-6 Astra",
      "input": 10,
      "output": 50,
      "blended": 20,
      "status": "verified",
      "sources": [
        "https://developers.openai.com/api/docs/models/gpt-6-astra"
      ]
    }
  ],
  "sources": [
    {
      "id": "s1",
      "title": "OpenAI pricing page, archived 2023-03-15 (GPT-4 8K context: $0.03/1K prompt, $0.06/1K completion)",
      "url": "https://web.archive.org/web/20230315033154/https://openai.com/pricing"
    },
    {
      "id": "s2",
      "title": "OpenAI DevDay announcement, archived 2023-11-06 ($0.01/1K input, $0.03/1K output)",
      "url": "https://web.archive.org/web/20231106184005/https://openai.com/blog/new-models-and-developer-products-announced-at-devday"
    },
    {
      "id": "s3",
      "title": "Mistral pricing page, archived 2024-02-26 (Mistral Large $8/$24 per 1M)",
      "url": "https://web.archive.org/web/20240226193838/https://docs.mistral.ai/platform/pricing/"
    },
    {
      "id": "s4",
      "title": "Anthropic, Introducing the next generation of Claude (Mar 4, 2024; Sonnet available in the API that day)",
      "url": "https://www.anthropic.com/news/claude-3-family"
    },
    {
      "id": "s5",
      "title": "Anthropic, Claude 3 Haiku release post (Mar 13, 2024); price from the Mar 4, 2024 Claude 3 announcement",
      "url": "https://www.anthropic.com/news/claude-3-haiku"
    },
    {
      "id": "s6",
      "title": "OpenAI, GPT-4o mini: advancing cost-efficient intelligence (Jul 18, 2024), archived",
      "url": "https://web.archive.org/web/20240718174226/https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/"
    },
    {
      "id": "s7",
      "title": "Google Developers Blog, Gemini 1.5 Flash price drop effective Aug 12, 2024 (prompts under 128K), archived",
      "url": "https://web.archive.org/web/20240815204013/https://developers.googleblog.com/en/gemini-15-flash-updates-google-ai-studio-gemini-api/"
    },
    {
      "id": "s8",
      "title": "Alibaba Cloud Model Studio models page (international), archived 2025-01-30 (qwen-turbo $0.00005/1K input, $0.0002/1K output)",
      "url": "https://web.archive.org/web/20250130085941/https://www.alibabacloud.com/help/en/model-studio/getting-started/models"
    },
    {
      "id": "s9",
      "title": "Alibaba Cloud Model Studio pricing, International (Singapore) deployment, prompts up to 32K tokens",
      "url": "https://www.alibabacloud.com/help/en/model-studio/model-pricing"
    },
    {
      "id": "s10",
      "title": "OpenAI API pricing page, archived 2024-05-13 (gpt-4o-2024-05-13 $5/$15)",
      "url": "https://web.archive.org/web/20240513221325/https://openai.com/api/pricing/"
    },
    {
      "id": "s11",
      "title": "OpenAI model page, GPT-4.5 Preview (snapshot gpt-4.5-preview-2025-02-27)",
      "url": "https://developers.openai.com/api/docs/models/gpt-4.5-preview"
    },
    {
      "id": "s12",
      "title": "OpenAI model page, GPT-5 (snapshot gpt-5-2025-08-07)",
      "url": "https://developers.openai.com/api/docs/models/gpt-5"
    },
    {
      "id": "s13",
      "title": "OpenAI model page, GPT-5.2 (snapshot gpt-5.2-2025-12-11)",
      "url": "https://developers.openai.com/api/docs/models/gpt-5.2"
    },
    {
      "id": "s14",
      "title": "OpenAI model page, GPT-5.4 (snapshot gpt-5.4-2026-03-05; prompts up to 272K)",
      "url": "https://developers.openai.com/api/docs/models/gpt-5.4"
    },
    {
      "id": "s15",
      "title": "OpenAI model page, GPT-5.5 (prompts up to 272K); API release Apr 24, 2026 per OpenAI changelog",
      "url": "https://developers.openai.com/api/docs/models/gpt-5.5"
    },
    {
      "id": "s16",
      "title": "OpenAI model page, GPT-6 Astra (prompts up to 272K); released Sep 3, 2026 per OpenAI changelog",
      "url": "https://developers.openai.com/api/docs/models/gpt-6-astra"
    },
    {
      "id": "s17",
      "title": "Epoch AI benchmark data, GPQA Diamond scores (downloaded 2026-10-01)",
      "url": "https://epoch.ai/data/benchmark_data.zip"
    }
  ]
}