AI Hub
Curated · Cross-linked · Editorial

The reference layer for frontier AI

One structured, hand-curated database of AI models, benchmarks, papers, and terms — cross-linked and kept current. The reference you bookmark and come back to, not a news feed.

660
Models
42
Benchmarks
73
Labs
107
Papers

Structured & cross-linked

Every model, paper, and term is connected. Jump from a benchmark to the models that lead it, or from a paper to the models it describes.

Editorial judgment

Hand-curated with a point of view on what actually matters — not an auto-generated dump of everything that shipped.

Always current

Pricing, context windows, and benchmarks sync from live sources, and new models are added as they ship.

Top models

Leading the intelligence index

A balanced average of each model’s per-category benchmark scores.

Full leaderboard →
  1. 1Sonar Reasoning Pro95.7
  2. 2R1 177695.4
  3. 3Qwen3.7 Max92.3
  4. 4Gemini 3.5 Flash92.2
  5. 5GPT-5.3-Codex91.5

Explore