---
slug: "inspect-ai"
name: "Inspect"
effect: "Defines and runs model and agent evaluations in Python from datasets, solvers and scorers"
category: "eval-observability"
tier: "rising"
url: "https://inspect.aisi.org.uk/"
github: "UKGovernmentBEIS/inspect_ai"
pricing: "open-source"
license: "MIT"
tags: ["evals", "python", "agents", "benchmarks", "sandbox", "cli"]
platforms: ["vscode"]
install_kind: "pip"
date_added: "2026-10-14"
rank: 31
rising_score: 31
stars: 3005
star_velocity: 2.8
velocity_basis: "since creation"
created_at: "2023-11-14T14:53:11Z"
pushed_at: "2026-10-15T01:13:29Z"
signals_status: "ok"
signals_fetched_at: "2026-10-15T05:02:42.515Z"
page: "https://agenttooldrop.com/tools/inspect-ai"
---

# Inspect

Defines and runs model and agent evaluations in Python from datasets, solvers and scorers

Category: Evals and observability. Tier: new and rising, rank 31 by new and rising score (31 of 100). Pricing: Open source. License: MIT.

Platforms: vscode.

## Why you would want it

Inspect, from the UK AI Security Institute and Meridian Labs, builds evaluations from composable datasets, solvers (from a single generate call to multi-turn agents with tools) and scorers, and runs them with `inspect eval`. It supports custom and MCP tools plus built-in bash, python, web browsing and computer tools, can evaluate external agents like Claude Code, Codex CLI and Gemini CLI, and sandboxes untrusted model code in Docker, Kubernetes, Modal and other systems. It ships over 200 pre-built benchmark evals, support for over 20 model providers, the Inspect View log viewer and a VS Code extension.

## Best for

- Benchmarking models and agents
- Agentic evals that run code in sandboxes
- Research and safety evaluation teams

## Use it when

You want to measure how a model or agent performs on a task suite, including agentic tasks with tools and sandboxed code, and compare results across providers.

## Skip it if

You need tracing of a live production app; Inspect is an evaluation framework, and evals are written in Python.

## Install

Install with pip:

```sh
pip install inspect-ai
```

Source: https://inspect.aisi.org.uk/

## Live signals

Fetched 2026-10-15T05:02:42.515Z (status: ok).

- Repository: UKGovernmentBEIS/inspect_ai
- Repository created: 2023-11-14
- Stars: 3,005 (GitHub's count; it cannot be authenticated)
- Stars per day, since repo creation: 2.8 (an average over the repo's life, not a growth rate)
- Forks: 811
- Open issues and PRs: 389
- License (GitHub): MIT
- Last push: 2026-10-15
- Latest release: 0.3.280 (2026-10-14)

## Links

- Website: https://inspect.aisi.org.uk/
- GitHub: https://github.com/UKGovernmentBEIS/inspect_ai
- Install instructions verified at: https://inspect.aisi.org.uk/
- Source for this entry: https://inspect.aisi.org.uk/
- AgentToolDrop page: https://agenttooldrop.com/tools/inspect-ai
- Full catalog (JSON): https://agenttooldrop.com/tools.json