{
 "name": "Rams Index",
 "edition": "models-v0.6.1",
 "season": "0.6.1",
 "url": "https://www.rams.ai/index",
 "method": "https://www.rams.ai/index/method",
 "updated": "2026-10-29",
 "license": "https://creativecommons.org/licenses/by/4.0/",
 "creator": "Rams (rams.ai)",
 "status": "Season 0.6.1. So far the judges are checked against 270 blind picks by one designer, Rams’ founder, and they agreed with him 63% of the time. Outside designers come next, with no date yet. Until then the Index is our judges' opinion: stable and repeatable, not yet validated by outside designers.",
 "index": {
  "formula": "logistic(mean Bradley-Terry log-strength of the model's pages), per brief",
  "range": [
   0,
   100
  ],
  "pairwise": "Inside each brief every page meets at least eight others. Claude Opus 5.5, GPT-6.1 Sol and Gemini 4 Argon each compare the two screenshots in both orders and pick the one a senior product designer would ship; a pick counts only when it names the same page both times. Bradley-Terry strengths per page, centered per brief; the Index is the chance a model's page beats an average page from the same brief, 50 = average. gradeSum is the earlier judges' score, kept for comparison.",
  "gradeSum": "Each judge gives seven 1-5 grades (hierarchy, typography, spacing, color, components, composure, craft), mean of desktop and phone, then passes. The two judges are z-normalized on their own page spread, averaged with equal weight, mapped back to the grade-sum scale, then averaged over briefs.",
  "interval": "95% bootstrap (B=4000), stratified by brief, resampling generations within a brief and judge passes within a page; the same random stream for every model, so differences pair by brief. A model with one generation per brief borrows the pooled between-generation sd from cells with 2+ generations (genNoiseSd).",
  "place": "Place on the board, 1 for the highest Rams Index (ties by the interval rank).",
  "rankRange": "Every place the paired 95% intervals allow: best is 1 plus the models clearly better, worst is the total minus the models clearly worse.",
  "costPerPage": "USD the model API charged to build one page, generation only, list price on run day.",
  "effort": "Every model runs at its provider default. No effort or thinking parameter is sent; a known default is named.",
  "broken": "A build is broken when it came back with no page (the token budget went to reasoning) or with a page cut off before it finished (the model hit the budget mid-file). Broken pages are counted and never scored. A model is ranked only with a scored page on every brief.",
  "criticals": "Share of scored pages whose first Rams code review has a critical finding.",
  "timeAndTokens": "Medians per page over every build, broken ones included. Reasoning tokens are what each API reports; null where it does not split them out."
 },
 "judges": {
  "score": "Claude Opus 5.5, GPT-6.1 Sol and Gemini 4 Argon",
  "cross": "GPT-5.6 Terra",
  "third": "OpenAI Decisions (planned, v4)"
 },
 "briefs": {
  "saas-landing": "Landing",
  "analytics-dashboard": "Dashboard",
  "settings-page": "Settings",
  "pricing-page": "Pricing",
  "article-page": "Article"
 },
 "viewports": [
  1440,
  390
 ],
 "dims": [
  "hierarchy",
  "typography",
  "spacing",
  "color",
  "components",
  "composure",
  "craft"
 ],
 "agreement": {
  "status": "pending",
  "readDate": "2026-11-06",
  "raters": 3,
  "tau": null,
  "tauLo": null,
  "tauHi": null,
  "ceilingTau": null
 },
 "models": [
  {
   "place": 1,
   "rankRange": {
    "best": 1,
    "worst": 2
   },
   "name": "GPT-6.1 Sol",
   "vendor": "OpenAI",
   "released": "2026-09",
   "slug": "gpt-6-1-sol",
   "id": "openai/gpt-6.1-sol",
   "url": "https://www.rams.ai/index/gpt-6-1-sol",
   "lab": "OpenAI",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0",
   "via": "OpenAI API",
   "servedBy": "OpenAI",
   "effort": "default",
   "medianSecondsPerPage": 101.5,
   "medianOutputTokens": 7218.5,
   "medianReasoningTokens": 366,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.8,
   "criticalPages": 8,
   "reviewedPages": 10,
   "index": 92,
   "lo": 90.3,
   "hi": 93.8,
   "gradeSum": 27.41,
   "usdPerPage": 0.15,
   "byBrief": {
    "analytics-dashboard": 87.1,
    "article-page": 94.5,
    "pricing-page": 93.3,
    "saas-landing": 95,
    "settings-page": 87
   },
   "dims": {
    "hierarchy": 3.8,
    "typography": 3.65,
    "spacing": 3.25,
    "color": 3.7,
    "components": 3.6,
    "composure": 3.6,
    "craft": 3.7
   },
   "terraIndex": 79.8,
   "visual": 88.75,
   "defectsPerPage": 5.2,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": true,
   "clearlyBehind": [],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 2,
   "rankRange": {
    "best": 1,
    "worst": 3
   },
   "name": "GPT-6 Astra",
   "vendor": "OpenAI",
   "released": "2026-09",
   "slug": "gpt-6-astra",
   "id": "openai/gpt-6-astra",
   "url": "https://www.rams.ai/index/gpt-6-astra",
   "lab": "OpenAI",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "OpenAI",
   "effort": "default",
   "medianSecondsPerPage": 113,
   "medianOutputTokens": 7682.5,
   "medianReasoningTokens": 501,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.9,
   "criticalPages": 9,
   "reviewedPages": 10,
   "index": 91,
   "lo": 88.7,
   "hi": 94.1,
   "gradeSum": 27.7,
   "usdPerPage": 0.428,
   "byBrief": {
    "analytics-dashboard": 82.6,
    "article-page": 95.7,
    "pricing-page": 92.6,
    "saas-landing": 96.6,
    "settings-page": 73.5
   },
   "dims": {
    "hierarchy": 3.85,
    "typography": 3.5,
    "spacing": 3.3,
    "color": 3.65,
    "components": 3.8,
    "composure": 3.8,
    "craft": 3.7
   },
   "terraIndex": 80.7,
   "visual": 90.65,
   "defectsPerPage": 6.5,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": true,
   "clearlyBehind": [],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 3,
   "rankRange": {
    "best": 3,
    "worst": 5
   },
   "name": "Grok 4.7",
   "vendor": "SpaceXSI",
   "released": "2026-09",
   "slug": "grok-4-7",
   "id": "x-ai/grok-4.7",
   "url": "https://www.rams.ai/index/grok-4-7",
   "lab": "SpaceXSI",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "SpaceXSI",
   "effort": "default",
   "medianSecondsPerPage": 283.5,
   "medianOutputTokens": 22134,
   "medianReasoningTokens": 13128,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0,
   "criticalPages": 0,
   "reviewedPages": 10,
   "index": 88.4,
   "lo": 85.9,
   "hi": 90.1,
   "gradeSum": 28.07,
   "usdPerPage": 0.132,
   "byBrief": {
    "analytics-dashboard": 92.5,
    "article-page": 64.6,
    "pricing-page": 97.6,
    "saas-landing": 92.8,
    "settings-page": 68.5
   },
   "dims": {
    "hierarchy": 3.7,
    "typography": 3.75,
    "spacing": 3.3,
    "color": 3.8,
    "components": 3.55,
    "composure": 3.65,
    "craft": 3.75
   },
   "terraIndex": 84.5,
   "visual": 88.6,
   "defectsPerPage": 8,
   "topDefects": [
    "Tap targets under 24px on the phone (8 of 10)",
    "Text under 12px (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": true,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-sol"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 4,
   "rankRange": {
    "best": 2,
    "worst": 5
   },
   "name": "Claude Fable 5.1",
   "vendor": "Anthropic",
   "released": "2026-08",
   "slug": "claude-fable-5-1",
   "id": "anthropic/claude-fable-5-1",
   "url": "https://www.rams.ai/index/claude-fable-5-1",
   "lab": "Anthropic",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0",
   "via": "Anthropic API",
   "servedBy": "Anthropic",
   "effort": "default (always thinks)",
   "medianSecondsPerPage": 180,
   "medianOutputTokens": 17284.5,
   "medianReasoningTokens": null,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.3,
   "criticalPages": 3,
   "reviewedPages": 10,
   "index": 88.3,
   "lo": 86.2,
   "hi": 89.9,
   "gradeSum": 25.97,
   "usdPerPage": 0.866,
   "byBrief": {
    "analytics-dashboard": 90.4,
    "article-page": 98.5,
    "pricing-page": 73,
    "saas-landing": 85.4,
    "settings-page": 70.5
   },
   "dims": {
    "hierarchy": 3.45,
    "typography": 3.35,
    "spacing": 3.1,
    "color": 3.35,
    "components": 3.35,
    "composure": 3.5,
    "craft": 3.4
   },
   "terraIndex": 77.1,
   "visual": 88.45,
   "defectsPerPage": 3.8,
   "topDefects": [
    "Tap targets under 24px on the phone (10 of 10)",
    "Text under 12px (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6.1-sol"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 5,
   "rankRange": {
    "best": 3,
    "worst": 5
   },
   "name": "GPT-6.1 Terra",
   "vendor": "OpenAI",
   "released": "2026-10",
   "slug": "gpt-6-1-terra",
   "id": "openai/gpt-6.1-terra",
   "url": "https://www.rams.ai/index/gpt-6-1-terra",
   "lab": "OpenAI",
   "weights": "closed",
   "weightsLink": null,
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "OpenAI",
   "effort": "default",
   "medianSecondsPerPage": 103.5,
   "medianOutputTokens": 7370,
   "medianReasoningTokens": 380.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.8,
   "criticalPages": 8,
   "reviewedPages": 10,
   "index": 86.1,
   "lo": 83.5,
   "hi": 90.1,
   "gradeSum": 27.45,
   "usdPerPage": 0.039,
   "byBrief": {
    "analytics-dashboard": 74.4,
    "article-page": 78.9,
    "pricing-page": 96.6,
    "saas-landing": 88.7,
    "settings-page": 78.9
   },
   "dims": {
    "hierarchy": 3.6,
    "typography": 3.6,
    "spacing": 3.25,
    "color": 3.65,
    "components": 3.6,
    "composure": 3.45,
    "craft": 3.65
   },
   "terraIndex": 82.9,
   "visual": 88.95,
   "defectsPerPage": 8.5,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Content wider than the phone (7 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": true,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-sol"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 6,
   "rankRange": {
    "best": 6,
    "worst": 11
   },
   "name": "GLM 5.3",
   "vendor": "Zhipu",
   "released": "2026-08",
   "slug": "glm-5-3",
   "id": "z-ai/glm-5.3",
   "url": "https://www.rams.ai/index/glm-5-3",
   "lab": "Zhipu",
   "weights": "open",
   "weightsLink": "https://huggingface.co/zai-org/GLM-5.3",
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Z.AI",
   "effort": "default",
   "medianSecondsPerPage": 969.5,
   "medianOutputTokens": 50912.5,
   "medianReasoningTokens": 37602.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.1,
   "criticalPages": 1,
   "reviewedPages": 10,
   "index": 76,
   "lo": 67.5,
   "hi": 78.5,
   "gradeSum": 29.38,
   "usdPerPage": 0.219,
   "byBrief": {
    "analytics-dashboard": 93.3,
    "article-page": 31.9,
    "pricing-page": 90.8,
    "saas-landing": 88.9,
    "settings-page": 37.9
   },
   "dims": {
    "hierarchy": 3.95,
    "typography": 3.9,
    "spacing": 3.3,
    "color": 3.9,
    "components": 3.8,
    "composure": 3.7,
    "craft": 3.8
   },
   "terraIndex": 91.1,
   "visual": 90.6,
   "defectsPerPage": 3.6,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (9 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 7,
   "rankRange": {
    "best": 6,
    "worst": 10
   },
   "name": "Claude Opus 5.5",
   "vendor": "Anthropic",
   "released": "2026-09",
   "slug": "claude-opus-5-5",
   "id": "anthropic/claude-opus-5-5",
   "url": "https://www.rams.ai/index/claude-opus-5-5",
   "lab": "Anthropic",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0",
   "via": "Anthropic API",
   "servedBy": "Anthropic",
   "effort": "default (medium)",
   "medianSecondsPerPage": 102,
   "medianOutputTokens": 13172.5,
   "medianReasoningTokens": null,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.1,
   "criticalPages": 1,
   "reviewedPages": 10,
   "index": 74.1,
   "lo": 70.9,
   "hi": 76.8,
   "gradeSum": 25.47,
   "usdPerPage": 0.276,
   "byBrief": {
    "analytics-dashboard": 79,
    "article-page": 91.4,
    "pricing-page": 60.7,
    "saas-landing": 49.9,
    "settings-page": 75.7
   },
   "dims": {
    "hierarchy": 3.6,
    "typography": 3.4,
    "spacing": 3.15,
    "color": 3.5,
    "components": 3.4,
    "composure": 3.45,
    "craft": 3.4
   },
   "terraIndex": 70.7,
   "visual": 86.5,
   "defectsPerPage": 3.2,
   "topDefects": [
    "Content wider than the phone (9 of 10)",
    "Tap targets under 24px on the phone (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 8,
   "rankRange": {
    "best": 6,
    "worst": 12
   },
   "name": "Claude Sonnet 5.5",
   "vendor": "Anthropic",
   "released": "2026-09",
   "slug": "claude-sonnet-5-5",
   "id": "anthropic/claude-sonnet-5.5",
   "url": "https://www.rams.ai/index/claude-sonnet-5-5",
   "lab": "Anthropic",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Anthropic",
   "effort": "default",
   "medianSecondsPerPage": 94.5,
   "medianOutputTokens": 13584.5,
   "medianReasoningTokens": 2720,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.1,
   "criticalPages": 1,
   "reviewedPages": 10,
   "index": 72.7,
   "lo": 66.6,
   "hi": 73.3,
   "gradeSum": 26.16,
   "usdPerPage": 0.145,
   "byBrief": {
    "analytics-dashboard": 94.7,
    "article-page": 27.2,
    "pricing-page": 70.5,
    "saas-landing": 63.3,
    "settings-page": 82.8
   },
   "dims": {
    "hierarchy": 3.75,
    "typography": 3.4,
    "spacing": 3.2,
    "color": 3.35,
    "components": 3.55,
    "composure": 3.75,
    "craft": 3.3
   },
   "terraIndex": 74.5,
   "visual": 90.15,
   "defectsPerPage": 2.6,
   "topDefects": [
    "Content wider than the phone (8 of 10)",
    "Tap targets under 24px on the phone (5 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 9,
   "rankRange": {
    "best": 6,
    "worst": 12
   },
   "name": "Mistral Large 4",
   "vendor": "Mistral",
   "released": "2026-10",
   "slug": "mistral-large-4-0",
   "id": "mistralai/mistral-large-4-0",
   "url": "https://www.rams.ai/index/mistral-large-4-0",
   "lab": "Mistral",
   "weights": "closed",
   "weightsLink": null,
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Mistral",
   "effort": "default",
   "medianSecondsPerPage": 671.5,
   "medianOutputTokens": 65298,
   "medianReasoningTokens": 53795.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.4,
   "criticalPages": 4,
   "reviewedPages": 10,
   "index": 72,
   "lo": 65.2,
   "hi": 74.8,
   "gradeSum": 28.27,
   "usdPerPage": 0.288,
   "byBrief": {
    "analytics-dashboard": 43.1,
    "article-page": 54.4,
    "pricing-page": 92.5,
    "saas-landing": 84.4,
    "settings-page": 64.9
   },
   "dims": {
    "hierarchy": 3.95,
    "typography": 3.85,
    "spacing": 3.15,
    "color": 4,
    "components": 3.5,
    "composure": 3.65,
    "craft": 3.75
   },
   "terraIndex": 84.3,
   "visual": 86.05,
   "defectsPerPage": 6.6,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (9 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 10,
   "rankRange": {
    "best": 6,
    "worst": 12
   },
   "name": "GPT-6 Luna",
   "vendor": "OpenAI",
   "released": "2026-09",
   "slug": "gpt-6-luna",
   "id": "openai/gpt-6-luna",
   "url": "https://www.rams.ai/index/gpt-6-luna",
   "lab": "OpenAI",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "OpenAI",
   "effort": "default",
   "medianSecondsPerPage": 81,
   "medianOutputTokens": 9163.5,
   "medianReasoningTokens": 797.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.7,
   "criticalPages": 7,
   "reviewedPages": 10,
   "index": 71.2,
   "lo": 67.4,
   "hi": 80.1,
   "gradeSum": 26.6,
   "usdPerPage": 0.005,
   "byBrief": {
    "analytics-dashboard": 81.8,
    "article-page": 17.7,
    "pricing-page": 76.3,
    "saas-landing": 82.7,
    "settings-page": 86.2
   },
   "dims": {
    "hierarchy": 3.65,
    "typography": 3.5,
    "spacing": 3.15,
    "color": 3.55,
    "components": 3.65,
    "composure": 3.3,
    "craft": 3.5
   },
   "terraIndex": 78.2,
   "visual": 88.3,
   "defectsPerPage": 5.3,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": true,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 11,
   "rankRange": {
    "best": 7,
    "worst": 13
   },
   "name": "Claude Haiku 5.5",
   "vendor": "Anthropic",
   "released": "2026-10",
   "slug": "claude-haiku-5-5",
   "id": "anthropic/claude-haiku-5.5",
   "url": "https://www.rams.ai/index/claude-haiku-5-5",
   "lab": "Anthropic",
   "weights": "closed",
   "weightsLink": null,
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Anthropic",
   "effort": "default",
   "medianSecondsPerPage": 51.5,
   "medianOutputTokens": 11313.5,
   "medianReasoningTokens": 651,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0,
   "criticalPages": 0,
   "reviewedPages": 10,
   "index": 70,
   "lo": 64.4,
   "hi": 73.7,
   "gradeSum": 25.27,
   "usdPerPage": 0.006,
   "byBrief": {
    "analytics-dashboard": 96.8,
    "article-page": 83.9,
    "pricing-page": 29.5,
    "saas-landing": 30.9,
    "settings-page": 70.1
   },
   "dims": {
    "hierarchy": 3.35,
    "typography": 3.25,
    "spacing": 3.1,
    "color": 3.35,
    "components": 3.3,
    "composure": 3.35,
    "craft": 3.3
   },
   "terraIndex": 73.8,
   "visual": 87.7,
   "defectsPerPage": 3.1,
   "topDefects": [
    "Tap targets under 24px on the phone (7 of 10)",
    "Content wider than the phone (6 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": true,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1",
    "anthropic/claude-opus-5-5"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 12,
   "rankRange": {
    "best": 8,
    "worst": 13
   },
   "name": "GPT-5.6 Sol",
   "vendor": "OpenAI",
   "released": "2026-06",
   "slug": "gpt-5-6-sol",
   "id": "openai/gpt-5.6-sol",
   "url": "https://www.rams.ai/index/gpt-5-6-sol",
   "lab": "OpenAI",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0",
   "via": "OpenAI API",
   "servedBy": "OpenAI",
   "effort": "default",
   "medianSecondsPerPage": 57,
   "medianOutputTokens": 8466,
   "medianReasoningTokens": 164,
   "broken": {
    "n": 0,
    "of": 15,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.33,
   "criticalPages": 5,
   "reviewedPages": 15,
   "index": 68.9,
   "lo": 62.8,
   "hi": 71.2,
   "gradeSum": 26.23,
   "usdPerPage": 0.175,
   "byBrief": {
    "analytics-dashboard": 54.1,
    "article-page": 76.9,
    "pricing-page": 27.9,
    "saas-landing": 49.4,
    "settings-page": 97.3
   },
   "dims": {
    "hierarchy": 3.4,
    "typography": 3.2,
    "spacing": 3.18,
    "color": 3.33,
    "components": 3.57,
    "composure": 3.43,
    "craft": 3.38
   },
   "terraIndex": 79.3,
   "visual": 88.57,
   "defectsPerPage": 4.7,
   "topDefects": [
    "Text under 12px (15 of 15)",
    "Tap targets under 24px on the phone (13 of 15)"
   ],
   "pages": 15,
   "passes": 20,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1",
    "anthropic/claude-opus-5-5"
   ],
   "provisional": [
    "1.3 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 13,
   "rankRange": {
    "best": 13,
    "worst": 15
   },
   "name": "Claude Fable 5",
   "vendor": "Anthropic",
   "released": "2026-06",
   "slug": "claude-fable-5",
   "id": "anthropic/claude-fable-5",
   "url": "https://www.rams.ai/index/claude-fable-5",
   "lab": "Anthropic",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0",
   "via": "Anthropic API",
   "servedBy": "Anthropic",
   "effort": "default (always thinks)",
   "medianSecondsPerPage": 99,
   "medianOutputTokens": 10314,
   "medianReasoningTokens": null,
   "broken": {
    "n": 0,
    "of": 15,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.2,
   "criticalPages": 3,
   "reviewedPages": 15,
   "index": 62.8,
   "lo": 57.6,
   "hi": 64.7,
   "gradeSum": 25.43,
   "usdPerPage": 0.532,
   "byBrief": {
    "analytics-dashboard": 71,
    "article-page": 66.7,
    "pricing-page": 67,
    "saas-landing": 15.6,
    "settings-page": 88.2
   },
   "dims": {
    "hierarchy": 3.33,
    "typography": 3.28,
    "spacing": 3.2,
    "color": 3.27,
    "components": 3.35,
    "composure": 3.38,
    "craft": 3.32
   },
   "terraIndex": 74.5,
   "visual": 88.7,
   "defectsPerPage": 3.2,
   "topDefects": [
    "Content wider than the phone (11 of 15)",
    "Text under 12px (10 of 15)"
   ],
   "pages": 15,
   "passes": 20,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "anthropic/claude-sonnet-5.5",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "anthropic/claude-opus-5-5",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.3 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 14,
   "rankRange": {
    "best": 11,
    "worst": 15
   },
   "name": "Kimi K3",
   "vendor": "Moonshot",
   "released": "2026-07",
   "slug": "kimi-k3",
   "id": "moonshotai/kimi-k3",
   "url": "https://www.rams.ai/index/kimi-k3",
   "lab": "Moonshot",
   "weights": "open",
   "weightsLink": "https://huggingface.co/moonshotai/Kimi-K3",
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Moonshot AI",
   "effort": "default",
   "medianSecondsPerPage": 633.5,
   "medianOutputTokens": 28376,
   "medianReasoningTokens": 16569.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.1,
   "criticalPages": 1,
   "reviewedPages": 10,
   "index": 62,
   "lo": 56.9,
   "hi": 67.3,
   "gradeSum": 26.72,
   "usdPerPage": 0.429,
   "byBrief": {
    "analytics-dashboard": 27.1,
    "article-page": 42.5,
    "pricing-page": 44.7,
    "saas-landing": 66.3,
    "settings-page": 96.4
   },
   "dims": {
    "hierarchy": 3.55,
    "typography": 3.5,
    "spacing": 3.15,
    "color": 3.5,
    "components": 3.45,
    "composure": 3.4,
    "craft": 3.55
   },
   "terraIndex": 80.4,
   "visual": 86.15,
   "defectsPerPage": 15.3,
   "topDefects": [
    "Text under 12px (9 of 10)",
    "Tap targets under 24px on the phone (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "anthropic/claude-sonnet-5.5",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-opus-5-5",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 15,
   "rankRange": {
    "best": 13,
    "worst": 16
   },
   "name": "Kimi K3.1 Pro",
   "vendor": "Moonshot",
   "released": "2026-10",
   "slug": "kimi-k3-1-pro",
   "id": "moonshotai/kimi-k3.1-pro",
   "url": "https://www.rams.ai/index/kimi-k3-1-pro",
   "lab": "Moonshot",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Moonshot AI",
   "effort": "default",
   "medianSecondsPerPage": 694,
   "medianOutputTokens": 29078,
   "medianReasoningTokens": 15447,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.6,
   "criticalPages": 6,
   "reviewedPages": 10,
   "index": 58.3,
   "lo": 55.2,
   "hi": 61,
   "gradeSum": 27.73,
   "usdPerPage": 0.317,
   "byBrief": {
    "analytics-dashboard": 45.6,
    "article-page": 78.6,
    "pricing-page": 30.6,
    "saas-landing": 50.1,
    "settings-page": 79.6
   },
   "dims": {
    "hierarchy": 3.6,
    "typography": 3.65,
    "spacing": 3.15,
    "color": 3.65,
    "components": 3.6,
    "composure": 3.75,
    "craft": 3.65
   },
   "terraIndex": 83.9,
   "visual": 90,
   "defectsPerPage": 3.7,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Content wider than the phone (6 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "anthropic/claude-sonnet-5.5",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "anthropic/claude-opus-5-5",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 16,
   "rankRange": {
    "best": 15,
    "worst": 16
   },
   "name": "GLM 5.3 Prime",
   "vendor": "Zhipu",
   "released": "2026-09",
   "slug": "glm-5-3-prime",
   "id": "z-ai/glm-5.3-prime",
   "url": "https://www.rams.ai/index/glm-5-3-prime",
   "lab": "Zhipu",
   "weights": "closed",
   "weightsLink": null,
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Alibaba",
   "effort": "default",
   "medianSecondsPerPage": 630,
   "medianOutputTokens": 56726.5,
   "medianReasoningTokens": 40870.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.2,
   "criticalPages": 2,
   "reviewedPages": 10,
   "index": 55.1,
   "lo": 48.8,
   "hi": 59.5,
   "gradeSum": 28.93,
   "usdPerPage": 0.496,
   "byBrief": {
    "analytics-dashboard": 61.4,
    "article-page": 49.8,
    "pricing-page": 81.8,
    "saas-landing": 70.7,
    "settings-page": 14
   },
   "dims": {
    "hierarchy": 3.9,
    "typography": 4.1,
    "spacing": 3.2,
    "color": 4.15,
    "components": 3.9,
    "composure": 3.65,
    "craft": 3.95
   },
   "terraIndex": 84.6,
   "visual": 90.6,
   "defectsPerPage": 6.4,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (10 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "anthropic/claude-sonnet-5.5",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "anthropic/claude-opus-5-5",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 17,
   "rankRange": {
    "best": 17,
    "worst": 18
   },
   "name": "DeepSeek V4 Pro",
   "vendor": "DeepSeek",
   "released": "2026-08",
   "slug": "deepseek-v4-pro-0813",
   "id": "deepseek/deepseek-v4-pro-0813",
   "url": "https://www.rams.ai/index/deepseek-v4-pro-0813",
   "lab": "DeepSeek",
   "weights": "open",
   "weightsLink": "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813",
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Alibaba",
   "effort": "default",
   "medianSecondsPerPage": 196,
   "medianOutputTokens": 15876,
   "medianReasoningTokens": 5827,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.1,
   "criticalPages": 1,
   "reviewedPages": 10,
   "index": 35.9,
   "lo": 31.3,
   "hi": 38.8,
   "gradeSum": 24.37,
   "usdPerPage": 0.032,
   "byBrief": {
    "analytics-dashboard": 21.6,
    "article-page": 37.4,
    "pricing-page": 45.8,
    "saas-landing": 27.8,
    "settings-page": 50.5
   },
   "dims": {
    "hierarchy": 3.25,
    "typography": 3.1,
    "spacing": 3.05,
    "color": 3.2,
    "components": 3.35,
    "composure": 3.15,
    "craft": 3.2
   },
   "terraIndex": 69.8,
   "visual": 87.3,
   "defectsPerPage": 3.2,
   "topDefects": [
    "Tap targets under 24px on the phone (7 of 10)",
    "Low-contrast text (6 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "anthropic/claude-opus-5-5",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 18,
   "rankRange": {
    "best": 18,
    "worst": 19
   },
   "name": "Claude Opus 4.8",
   "vendor": "Anthropic",
   "released": "2026-05",
   "slug": "claude-opus-4-8",
   "id": "anthropic/claude-opus-4-8",
   "url": "https://www.rams.ai/index/claude-opus-4-8",
   "lab": "Anthropic",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0",
   "via": "Anthropic API",
   "servedBy": "Anthropic",
   "effort": "default",
   "medianSecondsPerPage": 61,
   "medianOutputTokens": 6190,
   "medianReasoningTokens": null,
   "broken": {
    "n": 0,
    "of": 15,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.33,
   "criticalPages": 5,
   "reviewedPages": 15,
   "index": 32,
   "lo": 28.2,
   "hi": 34.7,
   "gradeSum": 24.02,
   "usdPerPage": 0.155,
   "byBrief": {
    "analytics-dashboard": 8.6,
    "article-page": 76.2,
    "pricing-page": 46,
    "saas-landing": 30.2,
    "settings-page": 17.2
   },
   "dims": {
    "hierarchy": 3.25,
    "typography": 3.1,
    "spacing": 3.05,
    "color": 3.05,
    "components": 3.02,
    "composure": 3.17,
    "craft": 3.17
   },
   "terraIndex": 69.5,
   "visual": 85.17,
   "defectsPerPage": 2.5,
   "topDefects": [
    "Content wider than the phone (9 of 15)",
    "Tap targets under 24px on the phone (8 of 15)"
   ],
   "pages": 15,
   "passes": 20,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.3 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 19,
   "rankRange": {
    "best": 17,
    "worst": 19
   },
   "name": "MiMo V2.6 Pro",
   "vendor": "Xiaomi",
   "released": "2026-09",
   "slug": "mimo-v2-6-pro",
   "id": "xiaomi/mimo-v2.6-pro",
   "url": "https://www.rams.ai/index/mimo-v2-6-pro",
   "lab": "Xiaomi",
   "weights": "open",
   "weightsLink": "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL",
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Xiaomi",
   "effort": "default",
   "medianSecondsPerPage": 338,
   "medianOutputTokens": 21638.5,
   "medianReasoningTokens": 4851,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0,
   "criticalPages": 0,
   "reviewedPages": 10,
   "index": 31.7,
   "lo": 26.6,
   "hi": 35.1,
   "gradeSum": 28.38,
   "usdPerPage": 0.021,
   "byBrief": {
    "analytics-dashboard": 44.6,
    "article-page": 35.9,
    "pricing-page": 48.5,
    "saas-landing": 62.5,
    "settings-page": 2.9
   },
   "dims": {
    "hierarchy": 3.85,
    "typography": 3.8,
    "spacing": 3.2,
    "color": 3.85,
    "components": 3.6,
    "composure": 3.75,
    "craft": 3.85
   },
   "terraIndex": 85,
   "visual": 89.1,
   "defectsPerPage": 12.6,
   "topDefects": [
    "Tap targets under 24px on the phone (10 of 10)",
    "Text under 12px (9 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "anthropic/claude-opus-5-5",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 20,
   "rankRange": {
    "best": 20,
    "worst": 20
   },
   "name": "Gemini 4 Argon",
   "vendor": "Google",
   "released": "2026-10",
   "slug": "gemini-4-argon",
   "id": "google/gemini-4-argon",
   "url": "https://www.rams.ai/index/gemini-4-argon",
   "lab": "Google",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Google AI Studio",
   "effort": "default",
   "medianSecondsPerPage": 170.5,
   "medianOutputTokens": 15682.5,
   "medianReasoningTokens": 0,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.5,
   "criticalPages": 5,
   "reviewedPages": 10,
   "index": 22.9,
   "lo": 19.7,
   "hi": 26.5,
   "gradeSum": 26.31,
   "usdPerPage": 0.149,
   "byBrief": {
    "analytics-dashboard": 19.5,
    "article-page": 27.5,
    "pricing-page": 33.5,
    "saas-landing": 10.9,
    "settings-page": 29.4
   },
   "dims": {
    "hierarchy": 3.7,
    "typography": 3.35,
    "spacing": 3.15,
    "color": 3.55,
    "components": 3.4,
    "composure": 3.65,
    "craft": 3.45
   },
   "terraIndex": 76.1,
   "visual": 86.4,
   "defectsPerPage": 3.1,
   "topDefects": [
    "Tap targets under 24px on the phone (8 of 10)",
    "Content wider than the phone (8 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "anthropic/claude-opus-4-8"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 21,
   "rankRange": {
    "best": 21,
    "worst": 23
   },
   "name": "Gemini 3.8 Flash",
   "vendor": "Google",
   "released": "2026-09",
   "slug": "gemini-3-8-flash",
   "id": "google/gemini-3.8-flash",
   "url": "https://www.rams.ai/index/gemini-3-8-flash",
   "lab": "Google",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Google AI Studio",
   "effort": "default",
   "medianSecondsPerPage": 93.5,
   "medianOutputTokens": 14409,
   "medianReasoningTokens": 1628.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.4,
   "criticalPages": 4,
   "reviewedPages": 10,
   "index": 11.5,
   "lo": 9.8,
   "hi": 13.9,
   "gradeSum": 24.51,
   "usdPerPage": 0.056,
   "byBrief": {
    "analytics-dashboard": 6.8,
    "article-page": 13.1,
    "pricing-page": 13.3,
    "saas-landing": 8.9,
    "settings-page": 18.6
   },
   "dims": {
    "hierarchy": 3.15,
    "typography": 3.25,
    "spacing": 3.1,
    "color": 3.35,
    "components": 3.35,
    "composure": 3.05,
    "craft": 3.3
   },
   "terraIndex": 69.6,
   "visual": 84.15,
   "defectsPerPage": 5,
   "topDefects": [
    "Tap targets under 24px on the phone (8 of 10)",
    "Content wider than the phone (7 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "google/gemini-4-argon",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "anthropic/claude-opus-4-8"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 22,
   "rankRange": {
    "best": 21,
    "worst": 24
   },
   "name": "MiniMax M3",
   "vendor": "MiniMax",
   "released": "2026-05",
   "slug": "minimax-m3",
   "id": "minimax/minimax-m3",
   "url": "https://www.rams.ai/index/minimax-m3",
   "lab": "MiniMax",
   "weights": "open",
   "weightsLink": "https://huggingface.co/MiniMaxAI/Minimax-M3",
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Minimax",
   "effort": "default",
   "medianSecondsPerPage": 144.5,
   "medianOutputTokens": 27248.5,
   "medianReasoningTokens": 11284.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.5,
   "criticalPages": 5,
   "reviewedPages": 10,
   "index": 10.8,
   "lo": 8.6,
   "hi": 16.7,
   "gradeSum": 26.44,
   "usdPerPage": 0.033,
   "byBrief": {
    "analytics-dashboard": 17.9,
    "article-page": 15.9,
    "pricing-page": 2.7,
    "saas-landing": 13.9,
    "settings-page": 12.4
   },
   "dims": {
    "hierarchy": 3.65,
    "typography": 3.6,
    "spacing": 3.2,
    "color": 3.65,
    "components": 3.85,
    "composure": 3.25,
    "craft": 3.55
   },
   "terraIndex": 74.5,
   "visual": 89.8,
   "defectsPerPage": 5.1,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (10 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "google/gemini-4-argon",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "anthropic/claude-opus-4-8"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 23,
   "rankRange": {
    "best": 21,
    "worst": 24
   },
   "name": "Qwen 3.8 Max",
   "vendor": "Alibaba",
   "released": "2026-09",
   "slug": "qwen3-8-max-0902",
   "id": "qwen/qwen3.8-max-0902",
   "url": "https://www.rams.ai/index/qwen3-8-max-0902",
   "lab": "Alibaba",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Alibaba",
   "effort": "default",
   "medianSecondsPerPage": 582.5,
   "medianOutputTokens": 24150.5,
   "medianReasoningTokens": 10048.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.5,
   "criticalPages": 5,
   "reviewedPages": 10,
   "index": 9,
   "lo": 7.6,
   "hi": 12.3,
   "gradeSum": 29.05,
   "usdPerPage": 0.149,
   "byBrief": {
    "analytics-dashboard": 14.1,
    "article-page": 2.7,
    "pricing-page": 29.2,
    "saas-landing": 30.2,
    "settings-page": 1.1
   },
   "dims": {
    "hierarchy": 4.05,
    "typography": 3.7,
    "spacing": 3.2,
    "color": 4,
    "components": 3.8,
    "composure": 3.65,
    "craft": 3.7
   },
   "terraIndex": 89.6,
   "visual": 85.85,
   "defectsPerPage": 12.9,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Content wider than the phone (10 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "google/gemini-4-argon",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "anthropic/claude-opus-4-8"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 24,
   "rankRange": {
    "best": 22,
    "worst": 24
   },
   "name": "Qwen 3.8 Max Prime",
   "vendor": "Alibaba",
   "released": "2026-09",
   "slug": "qwen3-8-max-prime",
   "id": "qwen/qwen3.8-max-prime",
   "url": "https://www.rams.ai/index/qwen3-8-max-prime",
   "lab": "Alibaba",
   "weights": "closed",
   "weightsLink": null,
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Alibaba",
   "effort": "default",
   "medianSecondsPerPage": 303,
   "medianOutputTokens": 25982.5,
   "medianReasoningTokens": 8903.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.5,
   "criticalPages": 5,
   "reviewedPages": 10,
   "index": 7.5,
   "lo": 7.1,
   "hi": 11,
   "gradeSum": 29.08,
   "usdPerPage": 0.342,
   "byBrief": {
    "analytics-dashboard": 21.6,
    "article-page": 7.5,
    "pricing-page": 1.8,
    "saas-landing": 30.8,
    "settings-page": 1.9
   },
   "dims": {
    "hierarchy": 4.25,
    "typography": 3.95,
    "spacing": 3.35,
    "color": 4,
    "components": 3.75,
    "composure": 3.85,
    "craft": 3.9
   },
   "terraIndex": 84.8,
   "visual": 89.4,
   "defectsPerPage": 14.7,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Content wider than the phone (9 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "google/gemini-4-argon",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "anthropic/claude-opus-4-8",
    "google/gemini-3.8-flash"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 25,
   "rankRange": {
    "best": 25,
    "worst": 26
   },
   "name": "Muse Spark Open",
   "vendor": "Meta",
   "released": "2026-10",
   "slug": "muse-spark-open",
   "id": "meta/muse-spark-open",
   "url": "https://www.rams.ai/index/muse-spark-open",
   "lab": "Meta",
   "weights": "open",
   "weightsLink": "https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B",
   "new": true,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Together",
   "effort": "default",
   "medianSecondsPerPage": 137,
   "medianOutputTokens": 13818,
   "medianReasoningTokens": 1635.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.4,
   "criticalPages": 4,
   "reviewedPages": 10,
   "index": 3.8,
   "lo": 2.9,
   "hi": 5.2,
   "gradeSum": 25.77,
   "usdPerPage": 0.062,
   "byBrief": {
    "analytics-dashboard": 2,
    "article-page": 2.3,
    "pricing-page": 4.4,
    "saas-landing": 4.7,
    "settings-page": 8
   },
   "dims": {
    "hierarchy": 3.45,
    "typography": 3.3,
    "spacing": 3,
    "color": 3.5,
    "components": 3.45,
    "composure": 3.25,
    "craft": 3.35
   },
   "terraIndex": 76.4,
   "visual": 86.9,
   "defectsPerPage": 6,
   "topDefects": [
    "Text under 12px (10 of 10)",
    "Tap targets under 24px on the phone (10 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "minimax/minimax-m3",
    "qwen/qwen3.8-max-prime",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "google/gemini-4-argon",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "qwen/qwen3.8-max-0902",
    "anthropic/claude-opus-4-8",
    "google/gemini-3.8-flash"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  },
  {
   "place": 26,
   "rankRange": {
    "best": 25,
    "worst": 26
   },
   "name": "Muse Wonder",
   "vendor": "Meta",
   "released": "2026-10",
   "slug": "muse-wonder",
   "id": "meta/muse-wonder",
   "url": "https://www.rams.ai/index/muse-wonder",
   "lab": "Meta",
   "weights": "closed",
   "weightsLink": null,
   "new": false,
   "season": "0.5",
   "via": "OpenRouter",
   "servedBy": "Meta",
   "effort": "default",
   "medianSecondsPerPage": 124.5,
   "medianOutputTokens": 12374.5,
   "medianReasoningTokens": 1068.5,
   "broken": {
    "n": 0,
    "of": 10,
    "noPage": 0,
    "truncated": 0
   },
   "criticalShare": 0.2,
   "criticalPages": 2,
   "reviewedPages": 10,
   "index": 2.6,
   "lo": 2.1,
   "hi": 4,
   "gradeSum": 25.98,
   "usdPerPage": 0.252,
   "byBrief": {
    "analytics-dashboard": 1.7,
    "article-page": 4.4,
    "pricing-page": 0.6,
    "saas-landing": 1.6,
    "settings-page": 16.6
   },
   "dims": {
    "hierarchy": 3.5,
    "typography": 3.35,
    "spacing": 3.15,
    "color": 3.55,
    "components": 3.5,
    "composure": 3.5,
    "craft": 3.4
   },
   "terraIndex": 74.8,
   "visual": 85.95,
   "defectsPerPage": 8.5,
   "topDefects": [
    "Low-contrast text (10 of 10)",
    "Text under 12px (10 of 10)"
   ],
   "pages": 10,
   "passes": 10,
   "frontier": false,
   "clearlyBehind": [
    "openai/gpt-6-astra",
    "z-ai/glm-5.3",
    "z-ai/glm-5.3-prime",
    "anthropic/claude-sonnet-5.5",
    "moonshotai/kimi-k3.1-pro",
    "minimax/minimax-m3",
    "qwen/qwen3.8-max-prime",
    "xiaomi/mimo-v2.6-pro",
    "openai/gpt-6.1-terra",
    "openai/gpt-6.1-sol",
    "anthropic/claude-fable-5",
    "x-ai/grok-4.7",
    "openai/gpt-5.6-sol",
    "anthropic/claude-fable-5-1",
    "openai/gpt-6-luna",
    "anthropic/claude-haiku-5.5",
    "deepseek/deepseek-v4-pro-0813",
    "anthropic/claude-opus-5-5",
    "google/gemini-4-argon",
    "moonshotai/kimi-k3",
    "mistralai/mistral-large-4-0",
    "qwen/qwen3.8-max-0902",
    "anthropic/claude-opus-4-8",
    "google/gemini-3.8-flash"
   ],
   "provisional": [
    "1.0 judge passes per page, protocol 2"
   ]
  }
 ],
 "unranked": [],
 "pairs": [
  {
   "a": "openai/gpt-6-astra",
   "b": "z-ai/glm-5.3",
   "diff": 15,
   "lo": 12.7,
   "hi": 23.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "z-ai/glm-5.3-prime",
   "diff": 35.9,
   "lo": 31,
   "hi": 44,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "anthropic/claude-sonnet-5.5",
   "diff": 18.3,
   "lo": 15.9,
   "hi": 25.2,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "moonshotai/kimi-k3.1-pro",
   "diff": 32.7,
   "lo": 29.1,
   "hi": 37,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "minimax/minimax-m3",
   "diff": 80.2,
   "lo": 75.2,
   "hi": 84,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "qwen/qwen3.8-max-prime",
   "diff": 83.5,
   "lo": 78.1,
   "hi": 86.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": 59.3,
   "lo": 54.8,
   "hi": 65.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "openai/gpt-6.1-terra",
   "diff": 4.9,
   "lo": 1.4,
   "hi": 10.2,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "openai/gpt-6.1-sol",
   "diff": -1,
   "lo": -3.2,
   "hi": 2.3,
   "separates": false
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "anthropic/claude-fable-5",
   "diff": 28.2,
   "lo": 25.6,
   "hi": 32.6,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "x-ai/grok-4.7",
   "diff": 2.6,
   "lo": 0.1,
   "hi": 6.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "openai/gpt-5.6-sol",
   "diff": 22.1,
   "lo": 19.6,
   "hi": 29.1,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "anthropic/claude-fable-5-1",
   "diff": 2.7,
   "lo": -1.1,
   "hi": 6.4,
   "separates": false
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "openai/gpt-6-luna",
   "diff": 19.8,
   "lo": 9.3,
   "hi": 25,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 21,
   "lo": 16.5,
   "hi": 27.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 55.1,
   "lo": 50.9,
   "hi": 60.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "meta/muse-spark-open",
   "diff": 87.2,
   "lo": 83.9,
   "hi": 90,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "anthropic/claude-opus-5-5",
   "diff": 16.9,
   "lo": 12.9,
   "hi": 21.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "google/gemini-4-argon",
   "diff": 68.1,
   "lo": 63,
   "hi": 73.1,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "moonshotai/kimi-k3",
   "diff": 29,
   "lo": 23.8,
   "hi": 36.9,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "mistralai/mistral-large-4-0",
   "diff": 19,
   "lo": 16.9,
   "hi": 27.9,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "meta/muse-wonder",
   "diff": 88.4,
   "lo": 85.8,
   "hi": 91.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 82,
   "lo": 79.2,
   "hi": 85.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "anthropic/claude-opus-4-8",
   "diff": 59,
   "lo": 54.5,
   "hi": 64.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6-astra",
   "b": "google/gemini-3.8-flash",
   "diff": 79.5,
   "lo": 76.5,
   "hi": 83.3,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "z-ai/glm-5.3-prime",
   "diff": 20.9,
   "lo": 13.3,
   "hi": 28.9,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "anthropic/claude-sonnet-5.5",
   "diff": 3.3,
   "lo": -3.7,
   "hi": 6.9,
   "separates": false
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "moonshotai/kimi-k3.1-pro",
   "diff": 17.7,
   "lo": 7.6,
   "hi": 20.8,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "minimax/minimax-m3",
   "diff": 65.2,
   "lo": 54.2,
   "hi": 67.7,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "qwen/qwen3.8-max-prime",
   "diff": 68.5,
   "lo": 58.5,
   "hi": 70.3,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": 44.3,
   "lo": 33.7,
   "hi": 47.6,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "openai/gpt-6.1-terra",
   "diff": -10.1,
   "lo": -20.4,
   "hi": -6.2,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "openai/gpt-6.1-sol",
   "diff": -16,
   "lo": -25.1,
   "hi": -13.9,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "anthropic/claude-fable-5",
   "diff": 13.2,
   "lo": 3.6,
   "hi": 19,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "x-ai/grok-4.7",
   "diff": -12.4,
   "lo": -21.3,
   "hi": -9.8,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "openai/gpt-5.6-sol",
   "diff": 7.1,
   "lo": 0.7,
   "hi": 10.4,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "anthropic/claude-fable-5-1",
   "diff": -12.3,
   "lo": -21.7,
   "hi": -9.7,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "openai/gpt-6-luna",
   "diff": 4.8,
   "lo": -12.4,
   "hi": 7.8,
   "separates": false
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 6,
   "lo": -4.9,
   "hi": 10.3,
   "separates": false
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 40.1,
   "lo": 30.3,
   "hi": 45.2,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "meta/muse-spark-open",
   "diff": 72.2,
   "lo": 62.8,
   "hi": 74.1,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "anthropic/claude-opus-5-5",
   "diff": 1.9,
   "lo": -8.2,
   "hi": 5.4,
   "separates": false
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "google/gemini-4-argon",
   "diff": 53.1,
   "lo": 42.1,
   "hi": 56.6,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "moonshotai/kimi-k3",
   "diff": 14,
   "lo": 4.4,
   "hi": 21,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "mistralai/mistral-large-4-0",
   "diff": 4,
   "lo": -5.9,
   "hi": 12.9,
   "separates": false
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "meta/muse-wonder",
   "diff": 73.4,
   "lo": 65,
   "hi": 75.9,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 67,
   "lo": 58.4,
   "hi": 70.1,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "anthropic/claude-opus-4-8",
   "diff": 44,
   "lo": 35.3,
   "hi": 48.2,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3",
   "b": "google/gemini-3.8-flash",
   "diff": 64.5,
   "lo": 56.1,
   "hi": 67.6,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "anthropic/claude-sonnet-5.5",
   "diff": -17.6,
   "lo": -22.6,
   "hi": -9.2,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "moonshotai/kimi-k3.1-pro",
   "diff": -3.2,
   "lo": -11,
   "hi": 2.7,
   "separates": false
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "minimax/minimax-m3",
   "diff": 44.3,
   "lo": 37.9,
   "hi": 50.5,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "qwen/qwen3.8-max-prime",
   "diff": 47.6,
   "lo": 40.8,
   "hi": 51.9,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": 23.4,
   "lo": 14.5,
   "hi": 30.2,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "openai/gpt-6.1-terra",
   "diff": -31,
   "lo": -38.6,
   "hi": -25.5,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "openai/gpt-6.1-sol",
   "diff": -36.9,
   "lo": -43.3,
   "hi": -31.4,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "anthropic/claude-fable-5",
   "diff": -7.7,
   "lo": -14.7,
   "hi": -0.3,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "x-ai/grok-4.7",
   "diff": -33.3,
   "lo": -39.7,
   "hi": -26.8,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "openai/gpt-5.6-sol",
   "diff": -13.8,
   "lo": -18.9,
   "hi": -8.1,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "anthropic/claude-fable-5-1",
   "diff": -33.2,
   "lo": -40.7,
   "hi": -27.8,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "openai/gpt-6-luna",
   "diff": -16.1,
   "lo": -30.9,
   "hi": -9.2,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -14.9,
   "lo": -23.7,
   "hi": -7.5,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 19.2,
   "lo": 10.9,
   "hi": 26.1,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "meta/muse-spark-open",
   "diff": 51.3,
   "lo": 44.1,
   "hi": 56.4,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "anthropic/claude-opus-5-5",
   "diff": -19,
   "lo": -27.6,
   "hi": -12.6,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "google/gemini-4-argon",
   "diff": 32.2,
   "lo": 23.9,
   "hi": 39.4,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "moonshotai/kimi-k3",
   "diff": -6.9,
   "lo": -12.6,
   "hi": -0.7,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "mistralai/mistral-large-4-0",
   "diff": -16.9,
   "lo": -22.1,
   "hi": -10,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "meta/muse-wonder",
   "diff": 52.5,
   "lo": 46.3,
   "hi": 56.5,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 46.1,
   "lo": 40.5,
   "hi": 50.5,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "anthropic/claude-opus-4-8",
   "diff": 23.1,
   "lo": 17.6,
   "hi": 30,
   "separates": true
  },
  {
   "a": "z-ai/glm-5.3-prime",
   "b": "google/gemini-3.8-flash",
   "diff": 43.6,
   "lo": 37.6,
   "hi": 47.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "moonshotai/kimi-k3.1-pro",
   "diff": 14.4,
   "lo": 6.7,
   "hi": 17.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "minimax/minimax-m3",
   "diff": 61.9,
   "lo": 50.1,
   "hi": 62.7,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "qwen/qwen3.8-max-prime",
   "diff": 65.2,
   "lo": 57.7,
   "hi": 65.3,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": 41,
   "lo": 35.5,
   "hi": 45.3,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "openai/gpt-6.1-terra",
   "diff": -13.4,
   "lo": -19.6,
   "hi": -10.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "openai/gpt-6.1-sol",
   "diff": -19.3,
   "lo": -24.5,
   "hi": -18.7,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "anthropic/claude-fable-5",
   "diff": 9.9,
   "lo": 3.9,
   "hi": 15.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "x-ai/grok-4.7",
   "diff": -15.7,
   "lo": -22.8,
   "hi": -14.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "openai/gpt-5.6-sol",
   "diff": 3.8,
   "lo": -1.8,
   "hi": 9,
   "separates": false
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "anthropic/claude-fable-5-1",
   "diff": -15.6,
   "lo": -19.8,
   "hi": -14.8,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "openai/gpt-6-luna",
   "diff": 1.5,
   "lo": -9.1,
   "hi": 4,
   "separates": false
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 2.7,
   "lo": -6.8,
   "hi": 7,
   "separates": false
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 36.8,
   "lo": 30.9,
   "hi": 40.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "meta/muse-spark-open",
   "diff": 68.9,
   "lo": 62.4,
   "hi": 69.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "anthropic/claude-opus-5-5",
   "diff": -1.4,
   "lo": -8.8,
   "hi": 0.9,
   "separates": false
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "google/gemini-4-argon",
   "diff": 49.8,
   "lo": 43.7,
   "hi": 51.9,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "moonshotai/kimi-k3",
   "diff": 10.7,
   "lo": 3.2,
   "hi": 15.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "mistralai/mistral-large-4-0",
   "diff": 0.7,
   "lo": -3.5,
   "hi": 6.3,
   "separates": false
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "meta/muse-wonder",
   "diff": 70.1,
   "lo": 64.3,
   "hi": 70.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 63.7,
   "lo": 54.6,
   "hi": 65.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "anthropic/claude-opus-4-8",
   "diff": 40.7,
   "lo": 34.5,
   "hi": 44.2,
   "separates": true
  },
  {
   "a": "anthropic/claude-sonnet-5.5",
   "b": "google/gemini-3.8-flash",
   "diff": 61.2,
   "lo": 53.8,
   "hi": 62.3,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "minimax/minimax-m3",
   "diff": 47.5,
   "lo": 40.2,
   "hi": 50,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "qwen/qwen3.8-max-prime",
   "diff": 50.8,
   "lo": 45.9,
   "hi": 53.5,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": 26.6,
   "lo": 23.1,
   "hi": 32.8,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "openai/gpt-6.1-terra",
   "diff": -27.8,
   "lo": -31.8,
   "hi": -23.8,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "openai/gpt-6.1-sol",
   "diff": -33.7,
   "lo": -38.6,
   "hi": -30.7,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "anthropic/claude-fable-5",
   "diff": -4.5,
   "lo": -8.8,
   "hi": 1.3,
   "separates": false
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "x-ai/grok-4.7",
   "diff": -30.1,
   "lo": -34.4,
   "hi": -27.1,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "openai/gpt-5.6-sol",
   "diff": -10.6,
   "lo": -15.4,
   "hi": -4.4,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "anthropic/claude-fable-5-1",
   "diff": -30,
   "lo": -33.4,
   "hi": -25.8,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "openai/gpt-6-luna",
   "diff": -12.9,
   "lo": -20.4,
   "hi": -7.4,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -11.7,
   "lo": -15.3,
   "hi": -4.5,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 22.4,
   "lo": 20.1,
   "hi": 27.1,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "meta/muse-spark-open",
   "diff": 54.5,
   "lo": 51,
   "hi": 56.6,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "anthropic/claude-opus-5-5",
   "diff": -15.8,
   "lo": -19.8,
   "hi": -11.9,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "google/gemini-4-argon",
   "diff": 35.4,
   "lo": 28.9,
   "hi": 39.7,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "moonshotai/kimi-k3",
   "diff": -3.7,
   "lo": -10.9,
   "hi": 3,
   "separates": false
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "mistralai/mistral-large-4-0",
   "diff": -13.7,
   "lo": -17.2,
   "hi": -6,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "meta/muse-wonder",
   "diff": 55.7,
   "lo": 51.8,
   "hi": 58.7,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 49.3,
   "lo": 45.7,
   "hi": 52.1,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "anthropic/claude-opus-4-8",
   "diff": 26.3,
   "lo": 23.3,
   "hi": 30.2,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3.1-pro",
   "b": "google/gemini-3.8-flash",
   "diff": 46.8,
   "lo": 42.9,
   "hi": 49.5,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "qwen/qwen3.8-max-prime",
   "diff": 3.3,
   "lo": -0.6,
   "hi": 8.4,
   "separates": false
  },
  {
   "a": "minimax/minimax-m3",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": -20.9,
   "lo": -24.5,
   "hi": -11.1,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "openai/gpt-6.1-terra",
   "diff": -75.3,
   "lo": -78.3,
   "hi": -68.7,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "openai/gpt-6.1-sol",
   "diff": -81.2,
   "lo": -83.3,
   "hi": -74.5,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "anthropic/claude-fable-5",
   "diff": -52,
   "lo": -53.8,
   "hi": -46,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "x-ai/grok-4.7",
   "diff": -77.6,
   "lo": -80.8,
   "hi": -72.4,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "openai/gpt-5.6-sol",
   "diff": -58.1,
   "lo": -62.4,
   "hi": -51,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "anthropic/claude-fable-5-1",
   "diff": -77.5,
   "lo": -79.8,
   "hi": -69.4,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "openai/gpt-6-luna",
   "diff": -60.4,
   "lo": -69.5,
   "hi": -51.5,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -59.2,
   "lo": -63,
   "hi": -50.5,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": -25.1,
   "lo": -28.6,
   "hi": -17.8,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "meta/muse-spark-open",
   "diff": 7,
   "lo": 4.8,
   "hi": 13.1,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "anthropic/claude-opus-5-5",
   "diff": -63.3,
   "lo": -67.2,
   "hi": -55.6,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "google/gemini-4-argon",
   "diff": -12.1,
   "lo": -15.9,
   "hi": -4.9,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "moonshotai/kimi-k3",
   "diff": -51.2,
   "lo": -56.9,
   "hi": -44.9,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "mistralai/mistral-large-4-0",
   "diff": -61.2,
   "lo": -64.1,
   "hi": -50.8,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "meta/muse-wonder",
   "diff": 8.2,
   "lo": 4.9,
   "hi": 14.3,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 1.8,
   "lo": -1.6,
   "hi": 7.8,
   "separates": false
  },
  {
   "a": "minimax/minimax-m3",
   "b": "anthropic/claude-opus-4-8",
   "diff": -21.2,
   "lo": -24,
   "hi": -14.7,
   "separates": true
  },
  {
   "a": "minimax/minimax-m3",
   "b": "google/gemini-3.8-flash",
   "diff": -0.7,
   "lo": -4.8,
   "hi": 5.3,
   "separates": false
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "xiaomi/mimo-v2.6-pro",
   "diff": -24.2,
   "lo": -27.4,
   "hi": -16.2,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "openai/gpt-6.1-terra",
   "diff": -78.6,
   "lo": -81.8,
   "hi": -73.2,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "openai/gpt-6.1-sol",
   "diff": -84.5,
   "lo": -86.5,
   "hi": -81,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "anthropic/claude-fable-5",
   "diff": -55.3,
   "lo": -57.2,
   "hi": -47.2,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "x-ai/grok-4.7",
   "diff": -80.9,
   "lo": -82.9,
   "hi": -77.6,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "openai/gpt-5.6-sol",
   "diff": -61.4,
   "lo": -63.2,
   "hi": -52.2,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "anthropic/claude-fable-5-1",
   "diff": -80.8,
   "lo": -82.1,
   "hi": -77,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "openai/gpt-6-luna",
   "diff": -63.7,
   "lo": -72.4,
   "hi": -59,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -62.5,
   "lo": -65.7,
   "hi": -54,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": -28.4,
   "lo": -31.1,
   "hi": -22.7,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "meta/muse-spark-open",
   "diff": 3.7,
   "lo": 2.6,
   "hi": 6.1,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "anthropic/claude-opus-5-5",
   "diff": -66.6,
   "lo": -69.7,
   "hi": -62.9,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "google/gemini-4-argon",
   "diff": -15.4,
   "lo": -19.3,
   "hi": -10.6,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "moonshotai/kimi-k3",
   "diff": -54.5,
   "lo": -58.6,
   "hi": -47.5,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "mistralai/mistral-large-4-0",
   "diff": -64.5,
   "lo": -64.8,
   "hi": -56.7,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "meta/muse-wonder",
   "diff": 4.9,
   "lo": 4.1,
   "hi": 8.4,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "qwen/qwen3.8-max-0902",
   "diff": -1.5,
   "lo": -4.2,
   "hi": 2.6,
   "separates": false
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "anthropic/claude-opus-4-8",
   "diff": -24.5,
   "lo": -25.1,
   "hi": -17.9,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-prime",
   "b": "google/gemini-3.8-flash",
   "diff": -4,
   "lo": -6.1,
   "hi": -0.9,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "openai/gpt-6.1-terra",
   "diff": -54.4,
   "lo": -61.2,
   "hi": -50.4,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "openai/gpt-6.1-sol",
   "diff": -60.3,
   "lo": -66.6,
   "hi": -56.4,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "anthropic/claude-fable-5",
   "diff": -31.1,
   "lo": -36.7,
   "hi": -26.5,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "x-ai/grok-4.7",
   "diff": -56.7,
   "lo": -62.6,
   "hi": -52.6,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "openai/gpt-5.6-sol",
   "diff": -37.2,
   "lo": -41.9,
   "hi": -32.7,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "anthropic/claude-fable-5-1",
   "diff": -56.6,
   "lo": -62.1,
   "hi": -53.3,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "openai/gpt-6-luna",
   "diff": -39.5,
   "lo": -46.9,
   "hi": -35.6,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -38.3,
   "lo": -45.6,
   "hi": -34.3,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": -4.2,
   "lo": -10,
   "hi": 0.8,
   "separates": false
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "meta/muse-spark-open",
   "diff": 27.9,
   "lo": 21.7,
   "hi": 30.6,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "anthropic/claude-opus-5-5",
   "diff": -42.4,
   "lo": -48.8,
   "hi": -38.6,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "google/gemini-4-argon",
   "diff": 8.8,
   "lo": 2.2,
   "hi": 13.4,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "moonshotai/kimi-k3",
   "diff": -30.3,
   "lo": -37.5,
   "hi": -22.9,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "mistralai/mistral-large-4-0",
   "diff": -40.3,
   "lo": -47.3,
   "hi": -31.2,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "meta/muse-wonder",
   "diff": 29.1,
   "lo": 24.3,
   "hi": 32.3,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 22.7,
   "lo": 15,
   "hi": 26.8,
   "separates": true
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "anthropic/claude-opus-4-8",
   "diff": -0.3,
   "lo": -5.3,
   "hi": 4.5,
   "separates": false
  },
  {
   "a": "xiaomi/mimo-v2.6-pro",
   "b": "google/gemini-3.8-flash",
   "diff": 20.2,
   "lo": 13.9,
   "hi": 23.8,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "openai/gpt-6.1-sol",
   "diff": -5.9,
   "lo": -9.2,
   "hi": -3.1,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "anthropic/claude-fable-5",
   "diff": 23.3,
   "lo": 19.9,
   "hi": 29.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "x-ai/grok-4.7",
   "diff": -2.3,
   "lo": -5.1,
   "hi": 1.2,
   "separates": false
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "openai/gpt-5.6-sol",
   "diff": 17.2,
   "lo": 13.6,
   "hi": 24.6,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "anthropic/claude-fable-5-1",
   "diff": -2.2,
   "lo": -5.5,
   "hi": 2.5,
   "separates": false
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "openai/gpt-6-luna",
   "diff": 14.9,
   "lo": 6.2,
   "hi": 21.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 16.1,
   "lo": 11.4,
   "hi": 24.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 50.2,
   "lo": 45.7,
   "hi": 56.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "meta/muse-spark-open",
   "diff": 82.3,
   "lo": 79,
   "hi": 86.9,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "anthropic/claude-opus-5-5",
   "diff": 12,
   "lo": 8,
   "hi": 18.5,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "google/gemini-4-argon",
   "diff": 63.2,
   "lo": 58.9,
   "hi": 67.1,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "moonshotai/kimi-k3",
   "diff": 24.1,
   "lo": 17.9,
   "hi": 29.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "mistralai/mistral-large-4-0",
   "diff": 14.1,
   "lo": 12.1,
   "hi": 21.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "meta/muse-wonder",
   "diff": 83.5,
   "lo": 80.9,
   "hi": 87.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 77.1,
   "lo": 73.5,
   "hi": 82.2,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "anthropic/claude-opus-4-8",
   "diff": 54.1,
   "lo": 49.4,
   "hi": 59.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-terra",
   "b": "google/gemini-3.8-flash",
   "diff": 74.6,
   "lo": 71.4,
   "hi": 78.8,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "anthropic/claude-fable-5",
   "diff": 29.2,
   "lo": 27.1,
   "hi": 35,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "x-ai/grok-4.7",
   "diff": 3.6,
   "lo": 1.1,
   "hi": 5.5,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "openai/gpt-5.6-sol",
   "diff": 23.1,
   "lo": 19.4,
   "hi": 30.6,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "anthropic/claude-fable-5-1",
   "diff": 3.7,
   "lo": 1.7,
   "hi": 6.6,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "openai/gpt-6-luna",
   "diff": 20.8,
   "lo": 11.3,
   "hi": 25.5,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 22,
   "lo": 17.6,
   "hi": 29,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 56.1,
   "lo": 53.6,
   "hi": 60.2,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "meta/muse-spark-open",
   "diff": 88.2,
   "lo": 86.2,
   "hi": 90.8,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "anthropic/claude-opus-5-5",
   "diff": 17.9,
   "lo": 14.4,
   "hi": 22.9,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "google/gemini-4-argon",
   "diff": 69.1,
   "lo": 64.6,
   "hi": 72,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "moonshotai/kimi-k3",
   "diff": 30,
   "lo": 24.5,
   "hi": 35.5,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "mistralai/mistral-large-4-0",
   "diff": 20,
   "lo": 18.3,
   "hi": 27.5,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "meta/muse-wonder",
   "diff": 89.4,
   "lo": 86.6,
   "hi": 91.3,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 83,
   "lo": 79.2,
   "hi": 85.6,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "anthropic/claude-opus-4-8",
   "diff": 60,
   "lo": 57.2,
   "hi": 65.1,
   "separates": true
  },
  {
   "a": "openai/gpt-6.1-sol",
   "b": "google/gemini-3.8-flash",
   "diff": 80.5,
   "lo": 76.9,
   "hi": 82.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "x-ai/grok-4.7",
   "diff": -25.6,
   "lo": -31.2,
   "hi": -23,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "openai/gpt-5.6-sol",
   "diff": -6.1,
   "lo": -12.3,
   "hi": -1.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "anthropic/claude-fable-5-1",
   "diff": -25.5,
   "lo": -31.9,
   "hi": -22.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "openai/gpt-6-luna",
   "diff": -8.4,
   "lo": -17.2,
   "hi": -3.7,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -7.2,
   "lo": -11.9,
   "hi": -1.7,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 26.9,
   "lo": 19.3,
   "hi": 31.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "meta/muse-spark-open",
   "diff": 59,
   "lo": 52.7,
   "hi": 61.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "anthropic/claude-opus-5-5",
   "diff": -11.3,
   "lo": -16.5,
   "hi": -6.9,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "google/gemini-4-argon",
   "diff": 39.9,
   "lo": 34.5,
   "hi": 43.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "moonshotai/kimi-k3",
   "diff": 0.8,
   "lo": -4.6,
   "hi": 5.4,
   "separates": false
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "mistralai/mistral-large-4-0",
   "diff": -9.2,
   "lo": -14.5,
   "hi": -3,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "meta/muse-wonder",
   "diff": 60.2,
   "lo": 54.7,
   "hi": 62.2,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 53.8,
   "lo": 47.9,
   "hi": 56.3,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "anthropic/claude-opus-4-8",
   "diff": 30.8,
   "lo": 23.4,
   "hi": 34.3,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5",
   "b": "google/gemini-3.8-flash",
   "diff": 51.3,
   "lo": 45.5,
   "hi": 54,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "openai/gpt-5.6-sol",
   "diff": 19.5,
   "lo": 16.2,
   "hi": 26.2,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "anthropic/claude-fable-5-1",
   "diff": 0.1,
   "lo": -2.3,
   "hi": 3.4,
   "separates": false
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "openai/gpt-6-luna",
   "diff": 17.2,
   "lo": 8,
   "hi": 22.3,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 18.4,
   "lo": 14.3,
   "hi": 25.2,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 52.5,
   "lo": 49.1,
   "hi": 56.9,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "meta/muse-spark-open",
   "diff": 84.6,
   "lo": 82.1,
   "hi": 86.7,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "anthropic/claude-opus-5-5",
   "diff": 14.3,
   "lo": 10.1,
   "hi": 18.9,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "google/gemini-4-argon",
   "diff": 65.5,
   "lo": 61.2,
   "hi": 69.5,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "moonshotai/kimi-k3",
   "diff": 26.4,
   "lo": 20.6,
   "hi": 31.8,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "mistralai/mistral-large-4-0",
   "diff": 16.4,
   "lo": 14,
   "hi": 23.5,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "meta/muse-wonder",
   "diff": 85.8,
   "lo": 82.1,
   "hi": 87.9,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 79.4,
   "lo": 76,
   "hi": 81.8,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "anthropic/claude-opus-4-8",
   "diff": 56.4,
   "lo": 53.8,
   "hi": 60.8,
   "separates": true
  },
  {
   "a": "x-ai/grok-4.7",
   "b": "google/gemini-3.8-flash",
   "diff": 76.9,
   "lo": 72.3,
   "hi": 78.5,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "anthropic/claude-fable-5-1",
   "diff": -19.4,
   "lo": -25.9,
   "hi": -16.2,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "openai/gpt-6-luna",
   "diff": -2.3,
   "lo": -13.3,
   "hi": 2.2,
   "separates": false
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "anthropic/claude-haiku-5.5",
   "diff": -1.1,
   "lo": -7.2,
   "hi": 3.2,
   "separates": false
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 33,
   "lo": 27.2,
   "hi": 37.5,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "meta/muse-spark-open",
   "diff": 65.1,
   "lo": 57.7,
   "hi": 67.6,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "anthropic/claude-opus-5-5",
   "diff": -5.2,
   "lo": -12.8,
   "hi": -2,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "google/gemini-4-argon",
   "diff": 46,
   "lo": 39.7,
   "hi": 51.2,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "moonshotai/kimi-k3",
   "diff": 6.9,
   "lo": -1,
   "hi": 12.4,
   "separates": false
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "mistralai/mistral-large-4-0",
   "diff": -3.1,
   "lo": -11,
   "hi": 3.1,
   "separates": false
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "meta/muse-wonder",
   "diff": 66.3,
   "lo": 60.3,
   "hi": 68.1,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 59.9,
   "lo": 52.8,
   "hi": 61.6,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "anthropic/claude-opus-4-8",
   "diff": 36.9,
   "lo": 31.9,
   "hi": 42,
   "separates": true
  },
  {
   "a": "openai/gpt-5.6-sol",
   "b": "google/gemini-3.8-flash",
   "diff": 57.4,
   "lo": 50.3,
   "hi": 59.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "openai/gpt-6-luna",
   "diff": 17.1,
   "lo": 9.3,
   "hi": 19.8,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 18.3,
   "lo": 12.7,
   "hi": 24.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 52.4,
   "lo": 49.6,
   "hi": 57.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "meta/muse-spark-open",
   "diff": 84.5,
   "lo": 81.9,
   "hi": 85.6,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "anthropic/claude-opus-5-5",
   "diff": 14.2,
   "lo": 10.4,
   "hi": 17.9,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "google/gemini-4-argon",
   "diff": 65.4,
   "lo": 61.4,
   "hi": 68.3,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "moonshotai/kimi-k3",
   "diff": 26.3,
   "lo": 21.8,
   "hi": 31.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "mistralai/mistral-large-4-0",
   "diff": 16.3,
   "lo": 13.1,
   "hi": 23.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "meta/muse-wonder",
   "diff": 85.7,
   "lo": 83.2,
   "hi": 87.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 79.3,
   "lo": 73.9,
   "hi": 81.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "anthropic/claude-opus-4-8",
   "diff": 56.3,
   "lo": 53.2,
   "hi": 60.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-fable-5-1",
   "b": "google/gemini-3.8-flash",
   "diff": 76.8,
   "lo": 73,
   "hi": 79.6,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "anthropic/claude-haiku-5.5",
   "diff": 1.2,
   "lo": -5.4,
   "hi": 9.4,
   "separates": false
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 35.3,
   "lo": 32.2,
   "hi": 44.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "meta/muse-spark-open",
   "diff": 67.4,
   "lo": 63.1,
   "hi": 75.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "anthropic/claude-opus-5-5",
   "diff": -2.9,
   "lo": -8,
   "hi": 4.6,
   "separates": false
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "google/gemini-4-argon",
   "diff": 48.3,
   "lo": 42.5,
   "hi": 55.1,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "moonshotai/kimi-k3",
   "diff": 9.2,
   "lo": 3.3,
   "hi": 19.2,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "mistralai/mistral-large-4-0",
   "diff": -0.8,
   "lo": -4.2,
   "hi": 9.5,
   "separates": false
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "meta/muse-wonder",
   "diff": 68.6,
   "lo": 64.3,
   "hi": 77.7,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 62.2,
   "lo": 55.8,
   "hi": 71.9,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "anthropic/claude-opus-4-8",
   "diff": 39.2,
   "lo": 35.6,
   "hi": 49.4,
   "separates": true
  },
  {
   "a": "openai/gpt-6-luna",
   "b": "google/gemini-3.8-flash",
   "diff": 59.7,
   "lo": 54.4,
   "hi": 68.8,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "deepseek/deepseek-v4-pro-0813",
   "diff": 34.1,
   "lo": 28.7,
   "hi": 39.1,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "meta/muse-spark-open",
   "diff": 66.2,
   "lo": 59.6,
   "hi": 69.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "anthropic/claude-opus-5-5",
   "diff": -4.1,
   "lo": -10.2,
   "hi": -0.8,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "google/gemini-4-argon",
   "diff": 47.1,
   "lo": 39.9,
   "hi": 51.9,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "moonshotai/kimi-k3",
   "diff": 8,
   "lo": -1.7,
   "hi": 15,
   "separates": false
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "mistralai/mistral-large-4-0",
   "diff": -2,
   "lo": -10.1,
   "hi": 7,
   "separates": false
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "meta/muse-wonder",
   "diff": 67.4,
   "lo": 62.2,
   "hi": 71.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 61,
   "lo": 55.7,
   "hi": 64.9,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "anthropic/claude-opus-4-8",
   "diff": 38,
   "lo": 31.7,
   "hi": 42.7,
   "separates": true
  },
  {
   "a": "anthropic/claude-haiku-5.5",
   "b": "google/gemini-3.8-flash",
   "diff": 58.5,
   "lo": 52.7,
   "hi": 62.4,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "meta/muse-spark-open",
   "diff": 32.1,
   "lo": 26.6,
   "hi": 34.7,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "anthropic/claude-opus-5-5",
   "diff": -38.2,
   "lo": -42.4,
   "hi": -35.2,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "google/gemini-4-argon",
   "diff": 13,
   "lo": 5.7,
   "hi": 15.9,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "moonshotai/kimi-k3",
   "diff": -26.1,
   "lo": -35,
   "hi": -20.6,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "mistralai/mistral-large-4-0",
   "diff": -36.1,
   "lo": -40.4,
   "hi": -30.4,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "meta/muse-wonder",
   "diff": 33.3,
   "lo": 28.4,
   "hi": 36.3,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 26.9,
   "lo": 22.2,
   "hi": 30.4,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "anthropic/claude-opus-4-8",
   "diff": 3.9,
   "lo": 0.3,
   "hi": 7.3,
   "separates": true
  },
  {
   "a": "deepseek/deepseek-v4-pro-0813",
   "b": "google/gemini-3.8-flash",
   "diff": 24.4,
   "lo": 20.2,
   "hi": 27.3,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "anthropic/claude-opus-5-5",
   "diff": -70.3,
   "lo": -72.6,
   "hi": -66.7,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "google/gemini-4-argon",
   "diff": -19.1,
   "lo": -23.5,
   "hi": -15.8,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "moonshotai/kimi-k3",
   "diff": -58.2,
   "lo": -62.9,
   "hi": -52.5,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "mistralai/mistral-large-4-0",
   "diff": -68.2,
   "lo": -70,
   "hi": -60.7,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "meta/muse-wonder",
   "diff": 1.2,
   "lo": -0.2,
   "hi": 2.7,
   "separates": false
  },
  {
   "a": "meta/muse-spark-open",
   "b": "qwen/qwen3.8-max-0902",
   "diff": -5.2,
   "lo": -8.1,
   "hi": -3.1,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "anthropic/claude-opus-4-8",
   "diff": -28.2,
   "lo": -29.6,
   "hi": -23.3,
   "separates": true
  },
  {
   "a": "meta/muse-spark-open",
   "b": "google/gemini-3.8-flash",
   "diff": -7.7,
   "lo": -9.9,
   "hi": -5.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "google/gemini-4-argon",
   "diff": 51.2,
   "lo": 44.4,
   "hi": 56.2,
   "separates": true
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "moonshotai/kimi-k3",
   "diff": 12.1,
   "lo": 4.6,
   "hi": 16.7,
   "separates": true
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "mistralai/mistral-large-4-0",
   "diff": 2.1,
   "lo": -0.5,
   "hi": 9,
   "separates": false
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "meta/muse-wonder",
   "diff": 71.5,
   "lo": 68.3,
   "hi": 74.5,
   "separates": true
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 65.1,
   "lo": 62.3,
   "hi": 68.3,
   "separates": true
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "anthropic/claude-opus-4-8",
   "diff": 42.1,
   "lo": 39.1,
   "hi": 47.4,
   "separates": true
  },
  {
   "a": "anthropic/claude-opus-5-5",
   "b": "google/gemini-3.8-flash",
   "diff": 62.6,
   "lo": 59,
   "hi": 65.4,
   "separates": true
  },
  {
   "a": "google/gemini-4-argon",
   "b": "moonshotai/kimi-k3",
   "diff": -39.1,
   "lo": -45,
   "hi": -34.5,
   "separates": true
  },
  {
   "a": "google/gemini-4-argon",
   "b": "mistralai/mistral-large-4-0",
   "diff": -49.1,
   "lo": -53,
   "hi": -43,
   "separates": true
  },
  {
   "a": "google/gemini-4-argon",
   "b": "meta/muse-wonder",
   "diff": 20.3,
   "lo": 15.8,
   "hi": 23.9,
   "separates": true
  },
  {
   "a": "google/gemini-4-argon",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 13.9,
   "lo": 9.1,
   "hi": 18.2,
   "separates": true
  },
  {
   "a": "google/gemini-4-argon",
   "b": "anthropic/claude-opus-4-8",
   "diff": -9.1,
   "lo": -13.3,
   "hi": -4,
   "separates": true
  },
  {
   "a": "google/gemini-4-argon",
   "b": "google/gemini-3.8-flash",
   "diff": 11.4,
   "lo": 6.1,
   "hi": 15.7,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3",
   "b": "mistralai/mistral-large-4-0",
   "diff": -10,
   "lo": -12.8,
   "hi": -3.8,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3",
   "b": "meta/muse-wonder",
   "diff": 59.4,
   "lo": 54.5,
   "hi": 64.5,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 53,
   "lo": 48,
   "hi": 57.9,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3",
   "b": "anthropic/claude-opus-4-8",
   "diff": 30,
   "lo": 24,
   "hi": 36.3,
   "separates": true
  },
  {
   "a": "moonshotai/kimi-k3",
   "b": "google/gemini-3.8-flash",
   "diff": 50.5,
   "lo": 45.5,
   "hi": 57.4,
   "separates": true
  },
  {
   "a": "mistralai/mistral-large-4-0",
   "b": "meta/muse-wonder",
   "diff": 69.4,
   "lo": 62.5,
   "hi": 72.5,
   "separates": true
  },
  {
   "a": "mistralai/mistral-large-4-0",
   "b": "qwen/qwen3.8-max-0902",
   "diff": 63,
   "lo": 54.4,
   "hi": 66,
   "separates": true
  },
  {
   "a": "mistralai/mistral-large-4-0",
   "b": "anthropic/claude-opus-4-8",
   "diff": 40,
   "lo": 34.5,
   "hi": 45.6,
   "separates": true
  },
  {
   "a": "mistralai/mistral-large-4-0",
   "b": "google/gemini-3.8-flash",
   "diff": 60.5,
   "lo": 53.1,
   "hi": 63.4,
   "separates": true
  },
  {
   "a": "meta/muse-wonder",
   "b": "qwen/qwen3.8-max-0902",
   "diff": -6.4,
   "lo": -10.1,
   "hi": -5,
   "separates": true
  },
  {
   "a": "meta/muse-wonder",
   "b": "anthropic/claude-opus-4-8",
   "diff": -29.4,
   "lo": -31.7,
   "hi": -24.8,
   "separates": true
  },
  {
   "a": "meta/muse-wonder",
   "b": "google/gemini-3.8-flash",
   "diff": -8.9,
   "lo": -11.4,
   "hi": -7,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-0902",
   "b": "anthropic/claude-opus-4-8",
   "diff": -23,
   "lo": -26.7,
   "hi": -18.3,
   "separates": true
  },
  {
   "a": "qwen/qwen3.8-max-0902",
   "b": "google/gemini-3.8-flash",
   "diff": -2.5,
   "lo": -4.5,
   "hi": 0.1,
   "separates": false
  },
  {
   "a": "anthropic/claude-opus-4-8",
   "b": "google/gemini-3.8-flash",
   "diff": 20.5,
   "lo": 15.3,
   "hi": 22.7,
   "separates": true
  }
 ],
 "judgesPanel": {
  "tau": {
   "models": {
    "tau": 0.74,
    "lo": 0.3,
    "hi": 0.76,
    "n": 26
   },
   "pages": {
    "tau": 0.63,
    "lo": 0.39,
    "hi": 0.68,
    "n": 275
   },
   "withinBrief": {
    "tau": 0.45,
    "lo": 0.28,
    "hi": 0.61,
    "n": 275
   }
  },
  "selfPreference": {
   "groups": {
    "Anthropic": [
     "anthropic/claude-sonnet-5.5",
     "anthropic/claude-fable-5",
     "anthropic/claude-fable-5-1",
     "anthropic/claude-haiku-5.5",
     "anthropic/claude-opus-5-5",
     "anthropic/claude-opus-4-8"
    ],
    "OpenAI": [
     "openai/gpt-6-astra",
     "openai/gpt-6.1-terra",
     "openai/gpt-6.1-sol",
     "openai/gpt-5.6-sol",
     "openai/gpt-6-luna"
    ]
   },
   "index": -0.5,
   "lo": -3.9,
   "hi": 3,
   "sd": -0.05,
   "sdLo": -0.37,
   "sdHi": 0.29,
   "claudeGap": -5.1,
   "terraGap": -4.6,
   "verdict": "none"
  }
 },
 "totals": {
  "pages": 275,
  "passes": 290,
  "models": 26,
  "unranked": 0,
  "brokenPages": 0
 }
}