Metatext

Metatext is a free, open directory that brings the AI ecosystem into one place. It lets researchers, engineers, and builders compare large language models by price, performance, and context window side by side. The data includes pricing, benchmarks, models, tools, and jobs, all in one searchable directory. It emphasizes openness and continual updates to reflect changes in the ecosystem.

The tool is designed for AI researchers, engineers, and builders who need to make fast, well-informed decisions about which model to ship, which dataset to train or evaluate on, and which tools or roles to plug into their workflows. It aggregates pricing, benchmarks, models, servers, skills, and job listings to support decision making.

Metatext pairs benchmark scores with price per million tokens, helping users find the best value rather than just the top score. It aggregates models, datasets, benchmarks, MCP servers, agent skills, and AI jobs in one place, aiming to streamline discovery and comparison.

Pricing, benchmarks, new models, MCP servers, agent skills, and job listings refresh continuously, with underlying data being open. The directory format supports filtering and exploring across providers and assets to support fast, informed decisions.

Use cases

• Researchers locate the best-value models for a task by comparing price and performance
• Engineers pick datasets to train or evaluate on from a central directory