A company building tools to understand increasingly capable AI
Create and share personal language model benchmarks, and compare state-of-the-art models
A tool to help you figure out how language models can help you in your job
Dynamic game-based benchmarks for language models. An intuitive way to assess model capabilities.
KraspAI was founded in Cambridge to make academic research into language model evaluation more widely accessible. The company's goal is to enable people to grasp AI’s potential, the upside and the risks — for the benefit of humanity.