{
  "name": "The LLM Conversation Cost Index",
  "description": "Cost of a standard 10-turn LLM conversation (system 2000, tools 1000, user 200/turn, assistant 500/turn, full history resent every turn), per model, with and without prompt caching. Snapshot of July 2026 list prices.",
  "source": "https://waynetools.rdkworld.com/llm-conversation-cost-index/",
  "license": "CC-BY 4.0",
  "attribution": "Data: The LLM Conversation Cost Index by waynetools.rdkworld.com, CC-BY",
  "snapshot_date": "2026-07-13",
  "methodology": {
    "system_tokens": 2000,
    "tools_tokens": 1000,
    "user_tokens_per_turn": 200,
    "assistant_tokens_per_turn": 500,
    "turns": 10,
    "input_tokens_total": 63500,
    "output_tokens_total": 5000,
    "cache_read_tokens_total": 54000,
    "cache_write_tokens_total": 9500,
    "formula": "input_turn_t = system + tools + (t-1)*(user+assistant) + user ; output never cached"
  },
  "models": [
    {
      "model": "GPT-4.1 nano",
      "provider": "OpenAI",
      "input_price_usd_per_1m": 0.1,
      "output_price_usd_per_1m": 0.4,
      "cache_read_multiplier": 0.25,
      "cache_write_multiplier": 1.0,
      "cost_per_conversation": 0.00835,
      "cost_per_1000_conversations": 8.35,
      "cost_with_caching_per_conversation": 0.0043,
      "cost_with_caching_per_1000_conversations": 4.3,
      "caching_savings_pct": 48.5,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Gemini 3 Flash",
      "provider": "Google",
      "input_price_usd_per_1m": 0.5,
      "output_price_usd_per_1m": 3.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 0.25,
      "cost_per_conversation": 0.04675,
      "cost_per_1000_conversations": 46.75,
      "cost_with_caching_per_conversation": 0.018887,
      "cost_with_caching_per_1000_conversations": 18.89,
      "caching_savings_pct": 59.6,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Claude Haiku 4.5",
      "provider": "Anthropic",
      "input_price_usd_per_1m": 1.0,
      "output_price_usd_per_1m": 5.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.0885,
      "cost_per_1000_conversations": 88.5,
      "cost_with_caching_per_conversation": 0.042275,
      "cost_with_caching_per_1000_conversations": 42.27,
      "caching_savings_pct": 52.2,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "GPT-5.6 Luna",
      "provider": "OpenAI",
      "input_price_usd_per_1m": 1.0,
      "output_price_usd_per_1m": 6.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.0935,
      "cost_per_1000_conversations": 93.5,
      "cost_with_caching_per_conversation": 0.047275,
      "cost_with_caching_per_1000_conversations": 47.27,
      "caching_savings_pct": 49.4,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Gemini 3.5 Flash",
      "provider": "Google",
      "input_price_usd_per_1m": 1.5,
      "output_price_usd_per_1m": 9.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 0.25,
      "cost_per_conversation": 0.14025,
      "cost_per_1000_conversations": 140.25,
      "cost_with_caching_per_conversation": 0.056662,
      "cost_with_caching_per_1000_conversations": 56.66,
      "caching_savings_pct": 59.6,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Claude Sonnet 5",
      "provider": "Anthropic",
      "input_price_usd_per_1m": 2.0,
      "output_price_usd_per_1m": 10.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.177,
      "cost_per_1000_conversations": 177.0,
      "cost_with_caching_per_conversation": 0.08455,
      "cost_with_caching_per_1000_conversations": 84.55,
      "caching_savings_pct": 52.2,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Gemini 3.1 Pro",
      "provider": "Google",
      "input_price_usd_per_1m": 2.0,
      "output_price_usd_per_1m": 12.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 0.25,
      "cost_per_conversation": 0.187,
      "cost_per_1000_conversations": 187.0,
      "cost_with_caching_per_conversation": 0.07555,
      "cost_with_caching_per_1000_conversations": 75.55,
      "caching_savings_pct": 59.6,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "GPT-5.6 Terra",
      "provider": "OpenAI",
      "input_price_usd_per_1m": 2.5,
      "output_price_usd_per_1m": 15.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.23375,
      "cost_per_1000_conversations": 233.75,
      "cost_with_caching_per_conversation": 0.118188,
      "cost_with_caching_per_1000_conversations": 118.19,
      "caching_savings_pct": 49.4,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Claude Opus 4.8",
      "provider": "Anthropic",
      "input_price_usd_per_1m": 5.0,
      "output_price_usd_per_1m": 25.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.4425,
      "cost_per_1000_conversations": 442.5,
      "cost_with_caching_per_conversation": 0.211375,
      "cost_with_caching_per_1000_conversations": 211.37,
      "caching_savings_pct": 52.2,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "GPT-5.6 Sol",
      "provider": "OpenAI",
      "input_price_usd_per_1m": 5.0,
      "output_price_usd_per_1m": 30.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.4675,
      "cost_per_1000_conversations": 467.5,
      "cost_with_caching_per_conversation": 0.236375,
      "cost_with_caching_per_1000_conversations": 236.38,
      "caching_savings_pct": 49.4,
      "snapshot_date": "2026-07-13"
    },
    {
      "model": "Claude Fable 5",
      "provider": "Anthropic",
      "input_price_usd_per_1m": 10.0,
      "output_price_usd_per_1m": 50.0,
      "cache_read_multiplier": 0.1,
      "cache_write_multiplier": 1.25,
      "cost_per_conversation": 0.885,
      "cost_per_1000_conversations": 885.0,
      "cost_with_caching_per_conversation": 0.42275,
      "cost_with_caching_per_1000_conversations": 422.75,
      "caching_savings_pct": 52.2,
      "snapshot_date": "2026-07-13"
    }
  ]
}