Skip to Content
Documentation

Welcome to OpenModels

OpenModels is an open-source registry for LLM models and inference providers. It helps developers discover, compare, and integrate AI models across multiple providers with standardized data and transparent pricing.

Key Features

Model Registry

A community-maintained, schema-validated registry of 223 LLM models as of October 24, 2026. Every model definition is machine-readable and version-controlled; hosted provider connections are tracked separately from canonical model records.

Provider Comparison

Compare 53 inference providers side-by-side — pricing, capabilities, context windows, and observed availability. Published prices preserve their original precision; unknown prices and rate limits are not presented as zero.

Real-Time Telemetry

Inspect observed provider reachability and inference latency. The current worker schedule runs health probes hourly and latency probes at 06:15 and 18:15 UTC, subject to credentials and provider policies. Missing or stale measurements display as no data, not 0% uptime or 0 ms latency. These synthetic checks are not a provider SLA or a guarantee that every model can serve authenticated requests.

Insights

AI industry analysis, provider comparisons, and benchmark reports published through our CMS-powered content platform. Stay informed about pricing changes, new model releases, and provider performance trends.

Open Source

The entire registry is open source and community-driven. Contribute new models, providers, and mappings through pull requests with automated schema validation.

What You’ll Find Here

  • Registry Repository — github.com/openmodelsrun/openmodels 
  • Coverage (October 24, 2026) — 223 models · 53 providers · 299 mappings; check /api/stats for current counts
  • Live Platform — openmodels.run  — browse, search, and compare models
  • API — api.openmodels.run  — public REST API, no auth required
  • Schema Definitions — Standardized YAML schemas for models, providers, and pricing
  • Automated Validation — All contributions are validated against JSON schemas on every PR