Awesome-AI4AI
Updated weekly — until humanity achieves AI4AI. 1462 curated papers across Benchmarks, Harness Design, and Model Design — from long-horizon agents to recursive
What is Awesome-AI4AI?
Updated weekly — until humanity achieves AI4AI. 1462 curated papers across Benchmarks, Harness Design, and Model Design — from long-horizon agents to recursive self-improvement. Citation rankings, GitHub stars, and source-verified weekly news.
Overview of Awesome-AI4AI
What it is
Awesome-AI4AI is a living catalog of papers, benchmarks, harness studies, and model studies focused on AI4AI — the concept of AI improving AI. It is updated weekly and grounded in a survey that covers long-horizon agents, recursive self-improvement, and related topics. The catalog currently contains 438 papers, 191 benchmarks, 224 harness studies, and 84 model studies, with overlaps. The original survey covers 223 papers.
Key features
- Weekly updates with new papers, releases, and news picks.
- Survey-grounded map from long-horizon agents to recursive self-improvement.
- Live rankings refreshed with maintainer updates, including top papers by average monthly citations and most-cited papers by year.
- Curated weekly news top 10, selected based on publisher authority, discussion volume, verifiable results, and durability.
- Community contributions integrated (e.g., ReCreate, Meta-Team, Agon, EvoRepair, etc.).
- Companion survey available online.
Use cases
Researchers and practitioners can use this repository to stay updated on the latest AI4AI research, track influential papers and benchmarks, and explore the landscape of long-horizon agents, harness design, and recursive self-improvement. It serves as a curated map for navigating the field.
Getting started
See the repository README for details on how to explore the collection, rankings, and weekly news.
License and pricing
MIT License. The project is open source.
Based on the project's public README and repository metadata.