Model Inclusion Policy

Modelindex is a curated registry of widely used AI foundation models. The goal is accuracy, traceability, and maintainability—not exhaustive coverage of every released model.

✅ A model is included only if it meets all of the following criteria: 

  • Published by a primary model developer organisation 
  • Officially documented by the provider 
  • Has a canonical model identifier (model code) 
  • Has clear lifecycle signals (active, deprecated, decommissioned) 
  • Has demonstrated production or broad industry usage 

🏢 Currently included providers 

  • OpenAI
  • Google
  • Anthropic
  • Meta
  • Mistral AI

These providers are included because they publish official model families, maintain documentation, and provide lifecycle visibility. 

🚫 Currently excluded 

The following are intentionally not indexed at this time: 

  • Community fine-tunes 
  • Third-party forks and repackaged models 
  • Experimental research releases 
  • Unofficial mirrors 
  • Long-tail open-source publishers 
  • Models without canonical model codes 
  • Models without official documentation pages 

📝 Notes 

Inclusion is based on provider authority and lifecycle transparency, not model quality. 
Exclusion does not imply a model is inferior—only that it falls outside the current scope of the registry.

The model inclusion scope may expand over time as providers mature their documentation and lifecycle practices. 

Scroll to Top