What AI Can My Computer Run?
Pick your graphics card or Mac and see which local models fit in its memory, and how much context each one leaves room for.
Hardware and model data can’t be loaded right now. Please try again in a few minutes.
Frequently asked questions
How do you decide whether a model fits?
We add the size of the model file, the memory its context needs, and a small allowance for the program running it, then compare that with the memory your graphics card or Mac can give it. The file size and the model’s design come from the publisher’s own pages; the allowance is our estimate.
What does “Partly on GPU” mean?
The model is too big for the graphics card alone, but the rest fits in your computer’s RAM. It will run, but much more slowly, because the part in RAM is worked through by the processor.
Why don’t you show how fast it will run?
Hardware makers and model publishers don’t publish tokens-per-second figures, so there is no source we could cite. We only show what we can back up.
Why does a longer context need more memory?
A model keeps working notes for every token in the conversation. The longer the context you load it with, the more memory those notes take, on top of the model file itself.