Meta's financial reports and analysis by financial news outlets

Meta's Llama 3 AI: 800 Tokens/Sec, Plans for 400B Parameters, Powered by Groq

Apr 20, 2024, 09:41 PM

Meta's new AI model, Llama 3, has been released with significant advancements in processing speed and capabilities. The model, developed by Meta and powered by Groq's AI chip, achieves an impressive 800 tokens per second, making it one of the fastest and most efficient AI models currently available. Llama 3 is available in various configurations, including 8B and 70B parameter models, with plans to train models up to 400B parameters. This model has been described as a game-changer for real-time applications and has outperformed competitors like Claude Opus and GPT-4 Turbo in speed, price, and quality metrics. The AI model is not only faster but also more cost-effective, offering performance close to GPT-4 at a fraction of the cost. Additionally, the release of Llama 3 has been termed a seismic event in the AI industry, enhancing AI assistance capabilities, though Meta's AI agents have shown some bizarre exchanges in social media interactions.