schema.jsonld

{
  "@context": "https://schema.org",
  "@type": "Dataset",
  "name": "ALL Bench Leaderboard 2026",
  "alternateName": ["ALL Bench", "ALLBench", "AI Benchmark Leaderboard 2026"],
  "description": "The only AI benchmark leaderboard covering LLM, VLM, Agent, Image, Video, and Music generation in a single unified view. 91 models cross-verified across 6 modalities with confidence badges. Features composite 5-axis scoring, interactive comparison tools, and downloadable intelligence reports.",
  "url": "https://huggingface.co/spaces/FINAL-Bench/all-bench-leaderboard",
  "sameAs": [
    "https://huggingface.co/datasets/FINAL-Bench/ALL-Bench-Leaderboard",
    "https://github.com/final-bench/ALL-Bench-Leaderboard"
  ],
  "license": "https://opensource.org/licenses/MIT",
  "version": "2.1",
  "datePublished": "2026-03-01",
  "dateModified": "2026-03-08",
  "creator": {
    "@type": "Organization",
    "name": "ALL Bench Team",
    "url": "https://huggingface.co/FINAL-Bench"
  },
  "keywords": [
    "AI benchmark", "LLM leaderboard", "GPT-5", "Claude", "Gemini",
    "VLM benchmark", "AI agent", "image generation", "video generation",
    "music generation", "MMLU-Pro", "GPQA", "ARC-AGI-2", "FINAL Bench",
    "metacognition", "multimodal AI", "AI evaluation", "benchmark comparison",
    "AI model ranking", "open source AI"
  ],
  "about": [
    {"@type": "Thing", "name": "Large Language Model"},
    {"@type": "Thing", "name": "Vision Language Model"},
    {"@type": "Thing", "name": "AI Benchmark"},
    {"@type": "Thing", "name": "Generative AI"},
    {"@type": "Thing", "name": "Metacognition"}
  ],
  "measurementTechnique": "Cross-verified benchmark aggregation with 3-tier confidence system",
  "variableMeasured": [
    {"@type": "PropertyValue", "name": "MMLU-Pro", "description": "57K expert-level multi-discipline questions"},
    {"@type": "PropertyValue", "name": "GPQA Diamond", "description": "PhD-level expert questions in science"},
    {"@type": "PropertyValue", "name": "AIME 2025", "description": "American Invitational Mathematics Examination"},
    {"@type": "PropertyValue", "name": "HLE", "description": "Humanity's Last Exam — 2500 expert-sourced questions"},
    {"@type": "PropertyValue", "name": "ARC-AGI-2", "description": "Abstract reasoning and novel pattern recognition"},
    {"@type": "PropertyValue", "name": "FINAL Bench Metacognitive", "description": "AI self-correction ability measurement"},
    {"@type": "PropertyValue", "name": "SWE-Pro", "description": "Software engineering benchmark by Scale AI"},
    {"@type": "PropertyValue", "name": "IFEval", "description": "Instruction following evaluation"},
    {"@type": "PropertyValue", "name": "LiveCodeBench", "description": "Continuously updated coding benchmark"}
  ],
  "distribution": [
    {
      "@type": "DataDownload",
      "encodingFormat": "application/json",
      "contentUrl": "https://huggingface.co/datasets/FINAL-Bench/ALL-Bench-Leaderboard/resolve/main/all_bench_leaderboard_v2.1.json",
      "name": "Unified JSON Dataset (75KB)"
    }
  ],
  "isPartOf": {
    "@type": "DataCatalog",
    "name": "Hugging Face Datasets",
    "url": "https://huggingface.co/datasets"
  },
  "funder": {
    "@type": "Organization",
    "name": "FINAL Bench"
  }
}