The best local LLM for every GPU memory bracket.

Choose your VRAM, use case, quantization, and context. We pick the biggest useful model that fits.