Benchmarks & Deep Dives · Free

Inspect

Inspect (inspect_ai) is an MIT-licensed open-source Python framework for LLM evaluations, created by the UK AI Security Institute and developed with Meridian Labs. Each eval is a Task built from a dataset, a solver and a scorer, with support for tool use, agents, MCP and sandboxed execution in Docker and other environments. It includes the Inspect View log viewer, a VS Code extension and a collection of more than 200 pre-built benchmark evals.