About

The definitive directory of the AI ecosystem.

Inference Index is a structured, editorially-maintained directory for everyone building on top of AI — founders, engineers, and analysts who need a single place to answer "which model?", "at what cost?", and "versus what?".

Models

Providers

Interfaces

Harnesses

Benchmarks

Plugins

Datasets

What we track

Seven layers of the stack.

The AI stack is not just "models". It is models, the providers that serve them, the interfaces that put them in front of humans, the harnesses that orchestrate them, the benchmarks that score them, the plugins they call, and the datasets that trained them. We index each layer separately and cross-link them.

How it works

One headline number, daily.

Prices are snapshotted daily and stored with full history. Benchmarks are updated as new results land. The Inference Index itself is computed from the top-scoring models on LMSYS ELO, weighted by published input/output pricing. Full methodology →

The news feed is ingested and scored every four hours from a curated list of labs, newsletters, and community sources. We publish a machine-readable RSS feed so you can plug it into whatever reader, Slack channel, or agent you like.

We're editorially independent. We do not take payment for placement or ranking. If we get something wrong, tell us.

Embed

Put the leaderboard on your site.

A lightweight iframe that shows the top models by LMSYS ELO, with the current blended price. Drop it in a blog post, docs page, or dashboard. Updates every hour. Themes: dark (default) or light.

<iframe
  src="https://inferenceindex.ai/embed/leaderboard?limit=5&theme=dark"
  width="520"
  height="360"
  frameborder="0"
  style="border: 0; max-width: 100%;"
  title="Inference Index leaderboard"
  loading="lazy"
></iframe>

Preview the widget →

Contact

Missing a model? Got a scoop?

The fastest way to reach us is email — hi@inferenceindex.ai. We also respond on Twitter / X. Corrections are triaged within 24 hours on weekdays.