Models.dev

TL;DR

A centralized, open-source repository for tracking AI model specs and pricing so you do not have to keep 40 browser tabs open.

Who is this actually for?

Engineers building LLM-powered apps who need to calculate unit economics or compare context windows across different providers.

The Good

• Open-source nature means anyone can submit a pull request to fix outdated info.
• Aggregates pricing and capabilities in one place, which is a massive headache to track manually.

The Catch (Potential Downsides)

The AI market moves at light speed; if this does not get constant community updates, the pricing data will be useless within weeks. It also lacks the nuance of real-world latency benchmarks which matter more than raw specs.