Llama 3.2 is a collection of large language models (LLMs) pretrained and fine-tuned in 1B and 3B sizes that are multilingual text only, and 11B and 90B sizes that take both text and image inputs and output text. https://www.llama.com/
inspired from https://manifold.markets/Moscow25/what-elo-will-openais-o1-model-get?play=true
1,000
3.00@LCBOB Llama 3.2 11b and 90b appear only on the vision leader board and receive 1034, and 1074. given that they are seemingly the same or very similar text-models to the similarly sized 3.1 models, just with an added vision component, I'm guessing they are only going to appearing on the vision board and if they appeared on the text board would receive similar rankings to their 3.1 counterparts, maxing out at 1211? The smaller models appear on the main board and get 1103, 1053.