CanIRun.ai

CanIRun.ai is a practical web tool that answers one specific question for developers and enthusiasts: can your computer run a particular AI model locally? You select your GPU model, available VRAM, and system RAM, and the tool shows you which models from providers like Meta, Google, Mistral AI, and OpenAI are compatible. It provides a detailed performance grade for each model, from 'Runs great' to 'Too heavy,' along with estimated memory usage and context window limits.

The tool works by comparing your hardware specifications against the requirements of hundreds of open-weight AI models. It shows not just binary compatibility but a nuanced performance score that considers quantization levels (Q2_K, Q4_K_M, etc.) and different model architectures. This helps you understand whether a model will run smoothly, be a tight fit, or simply won't work on your machine.

This is most useful for developers experimenting with local LLMs, researchers comparing model performance across hardware, and anyone building a PC for AI work. Instead of downloading multi-gigabyte models only to discover they won't run, you can use CanIRun.ai to make informed decisions about which models to try and what hardware upgrades might be needed.