Meta released Llama 3.1 405B, a model with 405 billion parameters, enhancing its problem-solving skills and competitive AI performance against leading proprietary models like OpenAI's GPT-4o.
Llama 3.1 405B is text-oriented and available for download or cloud use, powering chatbots on WhatsApp and Meta.ai.
Meta is exploring multimodality with Llama models that can handle images, videos, speech, but such capabilities are not yet publicly available.
The model was trained on a vast dataset of 15 trillion tokens to improve performance for text-based tasks in multiple languages.
[
Collection
]
[
|
...
]