{
  "id": 16872594,
  "title": "Hugging Face Official Benchmarks Map: 52 Boards, 1,131 Entries, and a New Retrieval Board",
  "url": "https://urgent.news/2026/10/28/hugging-face-official-benchmarks-map-52-boards-1-131-entries-and-a",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-28T08:38:49.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/quantid/hugging-face-official-benchmarks-map-52-boards-1131-entries-and-a-new-retrieval-board-24f"
  },
  "original_language": "en",
  "account": "The Hugging Face Official Benchmarks Map provides a comprehensive overview of all official benchmarks and leaderboards on the platform as of 28 October 2026. It contains 52 benchmarks, with 45 having at least one leaderboard entry. There are also 1,131 entries, with 438 distinct models from 104 organizations. One new field was added to the map: Retrieval, featuring BRIGHT, a reasoning-intensive retrieval benchmark. The Agents & terminal field leads with 17 leaderboards, accounting for the majority of entries, while Coding and Documents & OCR field together make up 19 leaderboards. The Retrieval field has the fewest leaderboards, with just 2. TIGER-Lab/MMLU-Pro, Idavidrein/gpqa, and cais/hle are the top three science & knowledge leaderboards, collectively holding 354 of the total entries (31.3%). The organization with the highest number of #1 spots is the map itself, with 20 organizations holding at least one top position. The scores on these leaderboards are self-reported and verified by Hugging Face, not by independent evaluation.",
  "summary": "TL;DR The Hugging Face Official Benchmarks Map reads every dataset tagged as an official benchmark on Hugging Face, and every leaderboard attached to it, straight from the public API each time the page opens. On 28 October 2026 (17:36 KST) it showed: Measure Value Official benchmarks 52 Benchmarks with a leaderboard 45 Leaderboard entries 1,131 Distinct models 438 Organizations 104 Since our 24…",
  "key_points": [
    "Hugging Face Official Benchmarks Map released on 28 October 2026",
    "52 benchmarks with 1,131 entries across 104 organizations",
    "Retrieval field added, featuring BRIGHT benchmark"
  ],
  "editors_take": "Expanding the map into specialized reasoning tasks signals a shift toward measuring complex retrieval capabilities within a landscape currently dominated by science and knowledge benchmarks.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}