Inspect
Open-source framework for LLM evaluations from the UK AI Security Institute, with epochs, exact scorers and standard error metrics.
0 members list Inspect, unchanged over the last 12 weeks, data as of 17 November 2026
Members who list Inspect in their Stack, by week.
Members who list Inspect
Be the first to add Inspect
Nobody on Stackness runs Inspect yet. Add it to your stack and you're the first member running it.
About Inspect
Inspect is an open-source framework for evaluating large language models, developed by the UK AI Security Institute. It provides composable building blocks including datasets, agents, tools, and scorers that allow developers to construct and reuse evaluations across a range of tasks such as coding, reasoning, knowledge assessment, and multi-modal understanding. The framework comes with over 200 pre-built evaluations ready to run on any model.
Inspect offers extensive tooling for evaluation monitoring and development, including a web-based visualization tool and a VS Code extension. It supports flexible tool calling with custom tools, built-in utilities like bash and web search, and agent evaluations with multi-agent capabilities. The framework includes a sandboxing system for running untrusted model code across various platforms such as Docker and Kubernetes. Inspect is written in Python and released under the MIT license.
Checked 11 November 2026.
Nobody on Stackness runs Inspect yet.
Sign up and add itPosts about Inspect
- How do you tell whether a hosted model got nerfed? The frozen eval panel, and what livenerf measures
A frozen eval panel is the only rerunnable way to tell whether a hosted model got worse. What livenerf measures on Opus 5.5, and how to size your own.