Meta's Chat-GPT: Llama 2 has been released as an open-source project.

On July 18, 2023, Meta (formerly known as Facebook) released the next-generation open-source large-scale language model (LLM) called 'LLama 2.' This model has been pre-trained and fine-tuned with 7 billion, 130 billion, and 700 billion parameters and was trained on 2 trillion tokens. It boasts twice the context length of its predecessor, 'LLama 1,' and was trained with over one million human annotations.

Meta stated, "We are harnessing the power of these large-scale language models," and announced that Llama 2, the latest version of LLama, is now available for personal use, creators, researchers, and businesses to responsibly experiment, innovate, and expand their ideas.

Mark Zuckerberg, CEO of Meta, announced via Facebook, "We are collaborating with Microsoft to release the next-generation open-source large-scale language model, Llama 2." He further stated, "Llama 2 will be available for research and commercial use for free."

Currently, the model can be downloaded and used with a simple registration, and for more detailed information about the model, one can refer to the research paper "LLama 2: Open Foundation and Fine-Tuned Chat Models," which was announced on the same day.