Meta today released Llama 3.1 405B, its largest and most capable large language model to date, which the social network claims can go toe-to-toe with OpenAI and Anthropic's top models.
Llama 3.1 better than GPT4 ?? OpenAI vs Meta with Llama 3.1 405B model
"Our experimental evaluation indicates that our flagship model is competitive with leading baseline models across a range of tasks, including GTP-4, GPT-4o and Claude 3.5 Sonnet," Meta boasted in an announcement, describing the neural network as "the world's largest and most capable openly available basic model." As you'd expect for an LLM, Llama 3.1 405B generates prose, chat responses and more from input messages.
Meta's Llama 3.1 405B, which was first teased at the launch of its eight- and 70-billion-parameter smaller siblings earlier this spring, was trained on more than 15 trillion tokens—viewing each as fragments of words, phrases, numbers, and punctuation marks—using of 16,000 Nvidia H100 GPUs.
In total, the Facebook giant says that training the model with 405 billion parameters required the equivalent of 30.84 million GPU hours and produced the equivalent of 11,390 tons of CO2 emissions.
