Meta is intensively developing a new open-source large-scale model that will compete with GPT-4. This model will have parameters several times larger than those of Llama 2. It is expected to be released next year and will support free commercial use. Meta's move is set to further disrupt the AI community, aiming to expand its influence through open-source models. This initiative also represents Meta's effort to bridge the gap between Llama 2 and GPT-4, hoping to get closer to the cutting edge of the industry technologically.