Look at models on huggingface, sorted by date, with keywords like "65b" (for llama 65b) or "ggml" (aka a pre quantized model for llama.cpp inference). https://

Look at models on huggingface, sorted by date, with keywords like "65b" (for llama 65b) or "ggml" (aka a pre quantized model for llama.cpp inference). https://huggingface.co/models?sort=modified&search=65b+ggml https://huggingface.co/models?sort=modified&search=65b+ggml Note that 33b llama models show up under "30b" or "33b" due to an old typo from Meta. I also look at what people are choosing to host here, though Kobold is mostly centered around the roleplaying crowd: https://lite.koboldai.net/# https://lite.koboldai.net/#