What happened
Museum of Models gathers answers produced by AI models, including systems that are no longer available. The archive preserves those responses exactly as they were produced, so that part of the history of these models is not lost.
The project began in 2024 and asks the same questions to different models. The collection currently holds 18,404 answers from 380 models to 100 questions.
Instead of building rankings, the platform lets users compare how the same tasks are answered in text, image, drawing or code. It also keeps incomplete, empty or faulty responses, and can be explored through an interactive globe and by searching for models or topics.
Key points
- The archive brings together answers from AI models, including those that have left circulation.
- The project keeps the content exactly as it was produced.
- The platform compares answers to the same questions, without rankings.
- The collection includes text, images, drawings, code and faulty responses.
Why it matters for your organisation
For companies and public organisations, this kind of archive helps show how AI models evolve and where they differ in the way they answer. That can support more informed choices about which systems to use and which tasks to trust them with.