PulseAugur
EN
LIVE 13:59:14

Hugging Face leaderboards: How benchmarks are populated and viewed

By PulseAugur Editorial · [3 sources] ·

Hugging Face maintains leaderboards for 48 official benchmarks, which are populated by model authors submitting results via .eval_results YAML files in their model repositories. These leaderboards are automatically assembled by Hugging Face and are not directly uploaded by benchmark owners. A significant portion of entries, around 30%, are not visible in the default view and require specific filtering. The "Agents and terminal" category hosts the most benchmarks, while "Science and knowledge" has the most entries, with MMLU-Pro, GPQA Diamond, and Humanity's Last Exam being among the most populated. AI

IMPACT Provides insight into how AI model performance is tracked and compared across various benchmarks.

RANK_REASON Article details the mechanics and data structure of Hugging Face's benchmark leaderboards, including how scores are submitted and displayed.

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 3 sources. How we write summaries →

Hugging Face leaderboards: How benchmarks are populated and viewed

How we ranked this

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
Article details the mechanics and data structure of Hugging Face's benchmark leaderboards, including how scores are submitted and displayed.
Source corroboration
3 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
product, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
18 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.
Coverage growth since scoring
+1 source(s) since last score
New sources have picked up this story since our last re-score. Score will update on the next scoring pass.

Full methodology in our editorial standards.

COVERAGE [3]

  1. X — Aravind Srinivas (Perplexity) TIER_1 English(EN) · AravSrinivas ·

    Perplexity wins on Hugging Face Decision Index (benchmark for decision models). Our model is open-weights and accessible to all!

    Perplexity wins on Hugging Face Decision Index (benchmark for decision models). Our model is open-weights and accessible to all!

  2. dev.to — LLM tag TIER_1 English(EN) · Ward Ed ·

    How Hugging Face Official Benchmark Leaderboards Actually Work: .eval_results YAML, the base_model Filter, and the 30% the Default View Hides

    <p><strong>TL;DR:</strong> Hugging Face currently tags 48 datasets as official benchmarks (<code>benchmark:official</code>). Their leaderboards are not uploaded by the benchmark owners: they are assembled automatically from small <code>.eval_results/*.yaml</code> files that model…

  3. dev.to — LLM tag TIER_1 English(EN) · AI OpenFree ·

    Hugging Face official benchmarks: the complete list (48) and how their leaderboards work

    <p><strong>Short answer:</strong> Hugging Face currently marks <strong>48 datasets as official benchmarks</strong> (October 4, 2026). Each has a leaderboard on its dataset page, built from <code>.eval_results</code> files that model repositories publish. Together they hold about …

RELATED ENTITIES

RELATED TOPICS