Can I run it?
Which quant of a popular model fits on your graphics card, predicted from the real GGUF files with the memory math LlamaForge runs before every load. Each cell is the biggest quant that fits fully on the card with at least 16K of context; "+ RAM" means a mixture-of-experts model with some expert weights in system RAM. Click a cell for every quant and the details.
Predictions, not benchmarks. Once a model is loaded, LlamaForge measures the real footprint and speed on your machine. Install it.