About ThinkLLM

ThinkLLM is a living, browsable knowledge graph about AI models. It's designed to help developers, researchers, and anyone curious about AI build practical intuition about the model landscape — which models exist, what they're good at, and how they compare.

How it works

AI agents continuously discover new models, gather metadata from publicly available sources, and synthesize original descriptions and capability scores. Every entry is reviewed before publication. The knowledge graph covers both open-weight models and commercial API models from major providers.

Model descriptions, capability scores, and glossary definitions are AI-generated content — original text produced by language models based on publicly available data. They are not copied from any source.

Data sources

  • Model metadata, model cards, and open-weight model discovery from HuggingFace Hub
  • Benchmark scores for open-weight models (IFEval, BBH, MATH, GPQA, MUSR, MMLU-PRO) from the Open LLM Leaderboard (Apache 2.0)
  • Human preference rankings from Chatbot Arena (Apache 2.0)
  • Commercial model pricing and benchmark data from Price Per Token
  • Commercial model announcements from provider RSS feeds and official documentation (Anthropic, OpenAI, Google DeepMind, Meta AI, Mistral, and others)
  • Web search context via Tavily (used transiently for paper discovery and model context)
  • Research papers linked from arXiv
  • Glossary definitions enriched from Wikipedia (CC BY-SA)

Contact

For corrections, takedown requests, or general inquiries: [email protected]

Disclaimer

Model names, product names, and trademarks referenced on this site belong to their respective owners. ThinkLLM is not affiliated with any model provider. Commercial model descriptions are editorial summaries based on publicly available information and are provided for informational and comparative purposes only.

Privacy PolicyTerms of Service