How jonathanchaveztamales/llmstats works
Project type and purpose
• This is an open-source, community-driven data repository of large language models (LLMs) and provider metadata.
• It does not implement a full frontend/backend stack itself, but it serves as the canonical data source for the interactive dashboard at llm-stats.com and any downstream consumers.
File structure
• Top-level directories:
• models/ – JSON files for each LLM organized by provider → model → model.json
• providers/ – JSON files describing each API provider → provider.json
• schemas/ – JSON Schema definitions (models-schema.json, providers-schema.json) and validator.js
• Tooling & configuration:
• package.json & package-lock.json – Node project to install dependencies (e.g., ajv for schema validation)
• .github/workflows/schema-validation.yml – CI pipeline that runs validator.js on all JSON data
• Contribution docs: CONTRIBUTING.md, LICENSE.md, PR templates, .gitignore
README for additional insights
• Describes the scope: model parameters, benchmarks, pricing, licensing
• Emphasizes the community contribution workflow and data quality (validation, multiple citations)
• Mentions an external interactive dashboard (llm-stats.com) that consumes this data