# AIBreakingNow

> Real-time tracker for AI model releases, benchmark results, and research papers. Auto-ranked by GitHub stars and citation counts.

## About

AIBreakingNow aggregates and ranks the latest developments in artificial intelligence:

- **AI Model Releases**: Sourced from HuggingFace Hub API, sorted by likes and recency
- **Research Papers**: From PapersWithCode, sorted by citations and publication date
- **GitHub Repositories**: Trending ML/AI repos sorted by star velocity

Data refreshes every hour via Incremental Static Regeneration (ISR).

## Pages

- `/` - English homepage with Top AI Today, trending models, papers, and repos
- `/es/` - Spanish (es-419)
- `/zh/` - Simplified Chinese (zh-CN)
- `/hi/` - Hindi (hi-IN)
- `/ar/` - Arabic (RTL)
- `/pt/` - Brazilian Portuguese
- `/fr/` - French (fr-FR)
- `/de/` - German (de-DE)
- `/ja/` - Japanese (ja-JP)
- `/ko/` - Korean (ko-KR)

## Feeds

- `/feed.xml` - RSS 2.0 feed of latest models, papers, and repos
- `/atom.xml` - Atom feed
- `/sitemap.xml` - Full XML sitemap
- `/sitemap-news.xml` - Google News sitemap

## Data Sources

All data is fetched from public APIs:

1. **HuggingFace Hub API**: `https://huggingface.co/api/models` - top text-generation models by likes
2. **PapersWithCode API**: `https://paperswithcode.com/api/v1/papers/` - latest ML research papers with citation counts
3. **GitHub Search API**: `https://api.github.com/search/repositories` - trending ML repositories by stars

## Pillar Content

The site includes a comprehensive evergreen guide: "Understanding AI Benchmarks & Model Releases" covering:
- MMLU, HumanEval, GSM8K, HellaSwag, ARC, TruthfulQA benchmark explanations
- How to evaluate new model releases
- GitHub stars as a market signal
- Model size vs. capability tradeoffs

## Technical

- Built with Next.js 14 (static export + ISR)
- 10 languages with full hreflang tags
- JSON-LD schemas: Organization, WebSite, ItemList, Article, FAQ, BreadcrumbList
- Open Graph and Twitter Card meta tags
- RSS 2.0 and Atom 1.0 feeds