{
 "video": "https://modelfatigue.news/v/semianalysis-subs/",
 "numbers": [
  {
   "id": "fee-200",
   "what": "monthly fee of the ChatGPT Pro 200 and Claude Max 20x plans",
   "value": 200,
   "shown": "$200",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "fee-100",
   "what": "monthly fee of the ChatGPT Pro 100 and Claude Max 5x plans",
   "value": 100,
   "shown": "$100",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "fee-20",
   "what": "monthly fee of the ChatGPT Plus and Claude Pro plans",
   "value": 20,
   "shown": "$20",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "fee-500",
   "what": "monthly fee of ChatGPT Pro 500",
   "value": 500,
   "shown": "$500",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "v-sol-200",
   "what": "API-equivalent value a month, GPT-6.1 Sol on ChatGPT Pro 200",
   "value": 2084,
   "shown": "$2,084",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "v-opus-200",
   "what": "API-equivalent value a month, Opus 5.5 on Claude Max 20x",
   "value": 11726,
   "shown": "$11,726",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-sol-200",
   "what": "API-equivalent value per dollar of fee, GPT-6.1 Sol on ChatGPT Pro 200 (SemiAnalysis’s multiple)",
   "value": 10.4,
   "shown": "10.4×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-opus-200",
   "what": "API-equivalent value per dollar of fee, Opus 5.5 on Claude Max 20x (SemiAnalysis’s multiple)",
   "value": 58.6,
   "shown": "58.6×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-sol-200",
   "what": "tokens a month, GPT-6.1 Sol on ChatGPT Pro 200",
   "value": 10200000000.0,
   "shown": "10.2B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "tpd-sol-200",
   "what": "tokens a month per dollar of fee, GPT-6.1 Sol on ChatGPT Pro 200",
   "value": 51000000.0,
   "shown": "51M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-opus-200",
   "what": "tokens a month, Opus 5.5 on Claude Max 20x",
   "value": 28600000000.0,
   "shown": "28.6B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "tpd-opus-200",
   "what": "tokens a month per dollar of fee, Opus 5.5 on Claude Max 20x",
   "value": 143000000.0,
   "shown": "143M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "v-sol-100",
   "what": "API-equivalent value a month, GPT-6.1 Sol on ChatGPT Pro 100",
   "value": 1055,
   "shown": "$1,055",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "v-opus-100",
   "what": "API-equivalent value a month, Opus 5.5 on Claude Max 5x",
   "value": 5725,
   "shown": "$5,725",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-sol-100",
   "what": "API-equivalent value per dollar of fee, GPT-6.1 Sol on ChatGPT Pro 100 (SemiAnalysis’s multiple)",
   "value": 10.6,
   "shown": "10.6×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-opus-100",
   "what": "API-equivalent value per dollar of fee, Opus 5.5 on Claude Max 5x (SemiAnalysis’s multiple)",
   "value": 57.3,
   "shown": "57.3×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-sol-100",
   "what": "tokens a month, GPT-6.1 Sol on ChatGPT Pro 100",
   "value": 5100000000.0,
   "shown": "5.1B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "tpd-sol-100",
   "what": "tokens a month per dollar of fee, GPT-6.1 Sol on ChatGPT Pro 100",
   "value": 51000000.0,
   "shown": "51M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-opus-100",
   "what": "tokens a month, Opus 5.5 on Claude Max 5x",
   "value": 14000000000.0,
   "shown": "14.0B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "tpd-opus-100",
   "what": "tokens a month per dollar of fee, Opus 5.5 on Claude Max 5x",
   "value": 140000000.0,
   "shown": "140M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "v-sol-20",
   "what": "API-equivalent value a month, GPT-6.1 Sol on ChatGPT Plus",
   "value": 211,
   "shown": "$211",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "v-opus-20",
   "what": "API-equivalent value a month, Opus 5.5 on Claude Pro",
   "value": 1178,
   "shown": "$1,178",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-sol-20",
   "what": "API-equivalent value per dollar of fee, GPT-6.1 Sol on ChatGPT Plus (SemiAnalysis’s multiple)",
   "value": 10.6,
   "shown": "10.6×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-opus-20",
   "what": "API-equivalent value per dollar of fee, Opus 5.5 on Claude Pro (SemiAnalysis’s multiple)",
   "value": 58.9,
   "shown": "58.9×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-sol-20",
   "what": "tokens a month, GPT-6.1 Sol on ChatGPT Plus",
   "value": 1000000000.0,
   "shown": "1.0B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "tpd-sol-20",
   "what": "tokens a month per dollar of fee, GPT-6.1 Sol on ChatGPT Plus",
   "value": 52000000.0,
   "shown": "52M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-opus-20",
   "what": "tokens a month, Opus 5.5 on Claude Pro",
   "value": 2900000000.0,
   "shown": "2.9B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "tpd-opus-20",
   "what": "tokens a month per dollar of fee, Opus 5.5 on Claude Pro",
   "value": 144000000.0,
   "shown": "144M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-opus",
   "what": "Opus 5.5, blended API list price per million tokens, agentic mix",
   "value": 0.4099,
   "shown": "$0.4099",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-sol",
   "what": "GPT-6.1 Sol, blended API list price per million tokens, agentic mix (printed rounded)",
   "value": 0.2,
   "shown": "$0.20",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their chart of the recent API price cuts (blended prices per million tokens in this mix, and Fable 5.1’s tokens on Max 20x)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-sol-old",
   "what": "GPT-6 Sol, blended API list price per million tokens, agentic mix, before GPT-6.1 Sol",
   "value": 0.3,
   "shown": "$0.30",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their chart of the recent API price cuts (blended prices per million tokens in this mix, and Fable 5.1’s tokens on Max 20x)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "t-fable-200",
   "what": "tokens a month, Fable 5.1 on Claude Max 20x (its half of the limit)",
   "value": 3200000000.0,
   "shown": "3.2B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their chart of the recent API price cuts (blended prices per million tokens in this mix, and Fable 5.1’s tokens on Max 20x)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-fable",
   "what": "Fable 5.1, blended API list price per million tokens, agentic mix",
   "value": 0.78,
   "shown": "$0.78",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their chart of the recent API price cuts (blended prices per million tokens in this mix, and Fable 5.1’s tokens on Max 20x)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "mix-in",
   "what": "SemiAnalysis’s agentic mix (their own usage in September): share of tokens that are new input",
   "value": 0.004,
   "shown": "0.4%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "mix-cr",
   "what": "SemiAnalysis’s agentic mix (their own usage in September): share of tokens that are cached input (re-reading context already sent)",
   "value": 0.966,
   "shown": "96.6%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "mix-cw",
   "what": "SemiAnalysis’s agentic mix (their own usage in September): share of tokens that are cache writes",
   "value": 0.026,
   "shown": "2.6%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "mix-out",
   "what": "SemiAnalysis’s agentic mix (their own usage in September): share of tokens that are output",
   "value": 0.003,
   "shown": "0.3%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "weeks",
   "what": "weeks in a month, as SemiAnalysis counts them",
   "value": 4.345,
   "shown": "4.345",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "stop-5",
   "what": "how narrow the range on a meter’s rate must be before SemiAnalysis stops counting steps (either way)",
   "value": 0.05,
   "shown": "5%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "cache-free",
   "what": "cache-read tokens after which an unmoved meter counts as free",
   "value": 500000000.0,
   "shown": "500M",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "essay-16k",
   "what": "longest essay asked for in each output call",
   "value": 16000,
   "shown": "16K",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "refused",
   "what": "the kind of prompt SemiAnalysis says models refuse, so they asked for essays instead",
   "value": "repeat SemiAnalysis 100,000 times",
   "shown": "repeat SemiAnalysis 100,000 times",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "ab-lower",
   "what": "how much lower the limits were on the one subscription in the provider’s A/B test (SemiAnalysis’s “~20%”)",
   "value": 0.2,
   "shown": "20%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, “Anthropic Subscriptions Offer 5x+ More Value Than OpenAI” (Andrew Megalaa, Max Kan, Dylan Patel, published 5 October 2026)",
   "url": "https://newsletter.semianalysis.com/p/anthropic-subscriptions-offer-5x",
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "f-astra-200",
   "what": "API-equivalent value a month, GPT-6 Astra on ChatGPT Pro 200",
   "value": 2897,
   "shown": "$2,897",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-astra-200",
   "what": "API-equivalent value per dollar of fee, GPT-6 Astra on ChatGPT Pro 200 (SemiAnalysis’s multiple)",
   "value": 14.5,
   "shown": "14.5×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "f-fable-200",
   "what": "API-equivalent value a month, Fable 5.1 on Claude Max 20x (at most half of the plan’s limit)",
   "value": 2485,
   "shown": "$2,485",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-fable-200",
   "what": "API-equivalent value per dollar of fee, Fable 5.1 on Claude Max 20x (SemiAnalysis’s multiple)",
   "value": 12.4,
   "shown": "12.4×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "f-astra-100",
   "what": "API-equivalent value a month, GPT-6 Astra on ChatGPT Pro 100",
   "value": 1322,
   "shown": "$1,322",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-astra-100",
   "what": "API-equivalent value per dollar of fee, GPT-6 Astra on ChatGPT Pro 100 (SemiAnalysis’s multiple)",
   "value": 13.2,
   "shown": "13.2×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "f-fable-100",
   "what": "API-equivalent value a month, Fable 5.1 on Claude Max 5x (at most half of the plan’s limit)",
   "value": 1273,
   "shown": "$1,273",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-fable-100",
   "what": "API-equivalent value per dollar of fee, Fable 5.1 on Claude Max 5x (SemiAnalysis’s multiple)",
   "value": 12.7,
   "shown": "12.7×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "f-astra-20",
   "what": "API-equivalent value a month, GPT-6 Astra on ChatGPT Plus",
   "value": 162,
   "shown": "$162",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-astra-20",
   "what": "API-equivalent value per dollar of fee, GPT-6 Astra on ChatGPT Plus (SemiAnalysis’s multiple)",
   "value": 8.1,
   "shown": "8.1×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “GPT-6 Astra and Claude Fable 5.1: API value by plan”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "half",
   "what": "share of a Max plan’s weekly limits Fable models may use at no extra cost",
   "value": 0.5,
   "shown": "50%",
   "whose": "Anthropic",
   "source": "Anthropic Help Center, “Claude Fable models on your plan” (dated 2 September 2026)",
   "url": "https://support.claude.com/en/articles/15424964-claude-fable-models-on-your-plan",
   "read": "2026-10-06T10:30",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "comp-sub",
   "what": "subscriptions’ share of Anthropic’s inference compute (SemiAnalysis’s rough estimate)",
   "value": 0.41700000000000004,
   "shown": "41.7%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "rev-sub",
   "what": "subscriptions’ share of Anthropic’s revenue (SemiAnalysis’s rough estimate)",
   "value": 0.1,
   "shown": "10.0%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "comp-api",
   "what": "API’s share of Anthropic’s inference compute (SemiAnalysis’s rough estimate)",
   "value": 0.583,
   "shown": "58.3%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "rev-api",
   "what": "API’s share of Anthropic’s revenue (SemiAnalysis’s rough estimate)",
   "value": 0.9,
   "shown": "90.0%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "util-20",
   "what": "the share of a plan’s limit SemiAnalysis calls a more realistic average use",
   "value": 0.2,
   "shown": "20%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "m-opus-20",
   "what": "Anthropic’s gross margin on Claude Max 20x if the subscriber uses 20% of what the plan allows on Opus 5.5 (SemiAnalysis’s estimate)",
   "value": 0.06,
   "shown": "6%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "m-fable-20",
   "what": "Anthropic’s gross margin on Claude Max 20x if the subscriber uses 20% of what the plan allows on Fable 5.1 (SemiAnalysis’s estimate)",
   "value": 0.8,
   "shown": "80%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "m-opus-100",
   "what": "Anthropic’s gross margin on Claude Max 20x if the subscriber uses all of it on Opus 5.5 (SemiAnalysis’s estimate)",
   "value": -3.69,
   "shown": "−369%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "m-fable-100",
   "what": "Anthropic’s gross margin on Claude Max 20x if the subscriber uses all Fable 5.1 may use (SemiAnalysis’s estimate)",
   "value": 0.01,
   "shown": "+1%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "api-margin",
   "what": "the API gross margin SemiAnalysis assumes for these estimates",
   "value": 0.92,
   "shown": "92%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-astra-old",
   "what": "API-equivalent value a month, GPT-6 Astra on ChatGPT Pro 200 before 29 September",
   "value": 5734,
   "shown": "$5,734",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-astra-pct",
   "what": "change in GPT-6 Astra’s value a month on ChatGPT Pro 200 at the 29 September cut",
   "value": -0.49,
   "shown": "−49%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-astra-tok-old",
   "what": "tokens a month, GPT-6 Astra on ChatGPT Pro 200 before 29 September",
   "value": 3800000000.0,
   "shown": "3.8B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-astra-tok-new",
   "what": "tokens a month, GPT-6 Astra on ChatGPT Pro 200 after 29 September",
   "value": 1900000000.0,
   "shown": "1.9B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-astra-tok-pct",
   "what": "change in GPT-6 Astra’s tokens a month on ChatGPT Pro 200 at the cut",
   "value": -0.49,
   "shown": "−49%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-sol-old",
   "what": "API-equivalent value a month, GPT-6 Sol on ChatGPT Pro 200 before 29 September",
   "value": 5904,
   "shown": "$5,904",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-sol-pct",
   "what": "change in value a month on ChatGPT Pro 200 at the cut, GPT-6 Sol before against GPT-6.1 Sol after",
   "value": -0.65,
   "shown": "−65%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-sol-tok-old",
   "what": "tokens a month, GPT-6 Sol on ChatGPT Pro 200 before 29 September",
   "value": 19600000000.0,
   "shown": "19.6B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "c-sol-tok-pct",
   "what": "change in tokens a month on ChatGPT Pro 200 at the cut, GPT-6 Sol before against GPT-6.1 Sol after",
   "value": -0.48,
   "shown": "−48%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-pct",
   "what": "ChatGPT Pro 500’s fee against the old Pro 200’s",
   "value": 1.5,
   "shown": "+150%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-astra",
   "what": "API-equivalent value a month, GPT-6 Astra on ChatGPT Pro 500",
   "value": 6955,
   "shown": "$6,955",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-astra-pct",
   "what": "GPT-6 Astra’s value a month on Pro 500 against the old Pro 200",
   "value": 0.21,
   "shown": "+21%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-astra-tok",
   "what": "tokens a month, GPT-6 Astra on ChatGPT Pro 500",
   "value": 4600000000.0,
   "shown": "4.6B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-astra-tok-pct",
   "what": "GPT-6 Astra’s tokens a month on Pro 500 against the old Pro 200",
   "value": 0.21,
   "shown": "+21%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-sol",
   "what": "API-equivalent value a month, GPT-6.1 Sol on ChatGPT Pro 500",
   "value": 5386,
   "shown": "$5,386",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-sol-pct",
   "what": "value a month on Pro 500 (GPT-6.1 Sol) against the old Pro 200 (GPT-6 Sol)",
   "value": -0.09,
   "shown": "−9%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-sol-tok",
   "what": "tokens a month, GPT-6.1 Sol on ChatGPT Pro 500",
   "value": 26300000000.0,
   "shown": "26.3B",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "p500-sol-tok-pct",
   "what": "tokens a month on Pro 500 (GPT-6.1 Sol) against the old Pro 200 (GPT-6 Sol)",
   "value": 0.34,
   "shown": "+34%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-astra-500",
   "what": "API-equivalent value per dollar of fee, GPT-6 Astra on ChatGPT Pro 500 (SemiAnalysis’s multiple)",
   "value": 13.9,
   "shown": "13.9×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT plans: API value multiple and implied gross margin” (the Pro 500 rows)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-sol-500",
   "what": "API-equivalent value per dollar of fee, GPT-6.1 Sol on ChatGPT Pro 500 (SemiAnalysis’s multiple)",
   "value": 10.8,
   "shown": "10.8×",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis, the same report: “ChatGPT plans: API value multiple and implied gross margin” (the Pro 500 rows)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "grand-from",
   "what": "the first day an active Pro 200 subscription qualifies for the old allowance",
   "value": "22 September",
   "shown": "22 September",
   "whose": "OpenAI",
   "source": "OpenAI Help Center, “About ChatGPT Pro tiers” (the 29 October grandfathering rule)",
   "url": "https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers",
   "read": "2026-10-21T01:45",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "grand-time",
   "what": "the time on 29 September by which a Pro 200 subscription had to be active to keep the old allowance",
   "value": "10 a.m. Pacific Time",
   "shown": "10 a.m. Pacific Time",
   "whose": "OpenAI",
   "source": "OpenAI Help Center, “About ChatGPT Pro tiers” (the 29 October grandfathering rule)",
   "url": "https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers",
   "read": "2026-10-21T01:45",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "ab-a",
   "what": "API-equivalent value a month, Opus 5.5 on Claude Max 20x, Account A in SemiAnalysis’s 22 October chart",
   "value": 11726,
   "shown": "$11,726",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis (@SemiAnalysis_) on X, 22 October 2026, and its chart “Anthropic appears to be A/B testing Claude Max 20x limits”",
   "url": "https://x.com/i/web/status/2113383965784502444",
   "read": "2026-10-26T07:41",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "ab-b",
   "what": "API-equivalent value a month, Opus 5.5 on Claude Max 20x, Account B in SemiAnalysis’s 22 October chart",
   "value": 9385,
   "shown": "$9,385",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis (@SemiAnalysis_) on X, 22 October 2026, and its chart “Anthropic appears to be A/B testing Claude Max 20x limits”",
   "url": "https://x.com/i/web/status/2113383965784502444",
   "read": "2026-10-26T07:41",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "ab-5h",
   "what": "Account B’s five-hour window against Account A’s, Opus 5.5 tokens (SemiAnalysis’s 22 October chart)",
   "value": 0.19,
   "shown": "+19%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis (@SemiAnalysis_) on X, 22 October 2026, and its chart “Anthropic appears to be A/B testing Claude Max 20x limits”",
   "url": "https://x.com/i/web/status/2113383965784502444",
   "read": "2026-10-26T07:41",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "ab-week",
   "what": "Account B’s weekly limit against Account A’s, Opus 5.5 tokens (SemiAnalysis’s 22 October chart)",
   "value": -0.2,
   "shown": "−20%",
   "whose": "SemiAnalysis",
   "source": "SemiAnalysis (@SemiAnalysis_) on X, 22 October 2026, and its chart “Anthropic appears to be A/B testing Claude Max 20x limits”",
   "url": "https://x.com/i/web/status/2113383965784502444",
   "read": "2026-10-26T07:41",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "cr-100",
   "what": "monthly API credits included with Claude Max 5x since 7 October",
   "value": 100,
   "shown": "$100",
   "whose": "Anthropic",
   "source": "Anthropic, “API credits for Max and Team plans”",
   "url": "https://platform.claude.com/docs/en/about-claude/api-credits-for-subscribers",
   "read": "2026-10-21T01:53",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "cr-200",
   "what": "monthly API credits included with Claude Max 20x since 7 October",
   "value": 200,
   "shown": "$200",
   "whose": "Anthropic",
   "source": "Anthropic, “API credits for Max and Team plans”",
   "url": "https://platform.claude.com/docs/en/about-claude/api-credits-for-subscribers",
   "read": "2026-10-21T01:53",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "sonnet-cr-old",
   "what": "Sonnet 5.5’s API price for prompt cache reads before 7 October, per million tokens",
   "value": 0.2,
   "shown": "$0.20",
   "whose": "Anthropic",
   "source": "Claude Platform, release notes (7 October 2026: Sonnet 5.5’s cache-read price)",
   "url": "https://platform.claude.com/docs/en/release-notes/overview",
   "read": "2026-10-21T01:45",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "sonnet-cr-new",
   "what": "Sonnet 5.5’s API price for prompt cache reads from 7 October, per million tokens",
   "value": 0.1,
   "shown": "$0.10",
   "whose": "Anthropic",
   "source": "Claude Platform, release notes (7 October 2026: Sonnet 5.5’s cache-read price)",
   "url": "https://platform.claude.com/docs/en/release-notes/overview",
   "read": "2026-10-21T01:45",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-v-200",
   "what": "API-equivalent value, Opus 5.5 on Claude Max 20x against GPT-6.1 Sol on ChatGPT Pro 200 (our arithmetic)",
   "value": 5.626679462571977,
   "shown": "5.6×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-t-200",
   "what": "tokens a month, Opus 5.5 on Claude Max 20x against GPT-6.1 Sol on ChatGPT Pro 200 (our arithmetic)",
   "value": 2.803921568627451,
   "shown": "2.8×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-v-100",
   "what": "API-equivalent value, Opus 5.5 on Claude Max 5x against GPT-6.1 Sol on ChatGPT Pro 100 (our arithmetic)",
   "value": 5.42654028436019,
   "shown": "5.4×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-t-100",
   "what": "tokens a month, Opus 5.5 on Claude Max 5x against GPT-6.1 Sol on ChatGPT Pro 100 (our arithmetic)",
   "value": 2.7450980392156863,
   "shown": "2.7×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-v-20",
   "what": "API-equivalent value, Opus 5.5 on Claude Pro against GPT-6.1 Sol on ChatGPT Plus (our arithmetic)",
   "value": 5.5829383886255926,
   "shown": "5.6×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-t-20",
   "what": "tokens a month, Opus 5.5 on Claude Pro against GPT-6.1 Sol on ChatGPT Plus (our arithmetic)",
   "value": 2.9,
   "shown": "2.9×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-opus-c",
   "what": "Opus 5.5, blended API price per million tokens, to the cent (our arithmetic)",
   "value": 0.4099,
   "shown": "$0.41",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-x",
   "what": "Opus 5.5’s blended price per token against GPT-6.1 Sol’s (our arithmetic)",
   "value": 2.0494999999999997,
   "shown": "2.05×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: their chart of the recent API price cuts (blended prices per million tokens in this mix, and Fable 5.1’s tokens on Max 20x)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "sol-derived",
   "what": "GPT-6.1 Sol’s value per million tokens on ChatGPT Pro 200 ($2,084 over 10.2 billion tokens) (our arithmetic)",
   "value": 0.20431372549019608,
   "shown": "$0.204",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “GPT-6.1 Sol and Claude Opus 5.5: API value by plan” and “… tokens by plan”, and their method chart’s blended price for Opus 5.5",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "price-x-fable",
   "what": "Fable 5.1’s blended price per token against Opus 5.5’s (our arithmetic)",
   "value": 1.9029031471090512,
   "shown": "1.9×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: their chart of the recent API price cuts (blended prices per million tokens in this mix, and Fable 5.1’s tokens on Max 20x)",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "comp-sub-r",
   "what": "subscriptions’ share of Anthropic’s inference compute, rounded as the video says it (our arithmetic)",
   "value": 0.41700000000000004,
   "shown": "42%",
   "whose": "SemiAnalysis (rough estimate), rounded by Model Fatigue",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "rev-sub-r",
   "what": "subscriptions’ share of Anthropic’s revenue, rounded as the video says it (our arithmetic)",
   "value": 0.1,
   "shown": "10%",
   "whose": "SemiAnalysis (rough estimate), rounded by Model Fatigue",
   "source": "SemiAnalysis, the same report: their rough estimates of Anthropic’s compute, revenue and subscription margins",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-astra-200-old",
   "what": "API-equivalent value per dollar of fee, GPT-6 Astra on ChatGPT Pro 200 before 29 September (our arithmetic)",
   "value": 28.67,
   "shown": "28.7×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "vf-sol-200-old",
   "what": "API-equivalent value per dollar of fee, GPT-6 Sol on ChatGPT Pro 200 before 29 September (our arithmetic)",
   "value": 29.52,
   "shown": "29.5×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis, the same report: “ChatGPT Pro 200: usage limits before and after September 29th” and “ChatGPT Pro 500 vs. Pro 200 before the cut”",
   "url": null,
   "read": "2026-10-06T09:32",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  },
  {
   "id": "x-v-ab",
   "what": "API-equivalent value, Opus 5.5 on Account B’s Claude Max 20x (22 October) against GPT-6.1 Sol on ChatGPT Pro 200 (5 October) (our arithmetic)",
   "value": 4.503358925143954,
   "shown": "4.5×",
   "whose": "Model Fatigue, arithmetic on SemiAnalysis’s figures",
   "source": "SemiAnalysis (@SemiAnalysis_) on X, 22 October 2026, and its chart “Anthropic appears to be A/B testing Claude Max 20x limits”",
   "url": "https://x.com/i/web/status/2113383965784502444",
   "read": "2026-10-26T07:41",
   "moves": false,
   "caveat": null,
   "now": null,
   "now_shown": null,
   "now_read": null
  }
 ]
}