Quantcast
Channel: Internet – MySmartPrice
Viewing all articles
Browse latest Browse all 306

Meta Unveils Llama 3.1, World’s Largest Open-Source AI Model

$
0
0
Highlights
  • Meta has released its powerful and advanced Llama 3.1 AI model and kept it open-source for all developers.
  • The new AI model also supports Hinglish, where users can speak to the chatbot in Hindi by using English (Roman) alphabets.

Meta has launched its Llama 3.1 AI model with improved reasoning and logical capabilities. The highlight of this large language model (LLM) is its open-source nature. Meta says Llama 3.1 is the world’s largest open-source AI model that can compete against the closed-system AI models from OpenAI, Claude, and others. Here are the details.

Meta Llama 3.1 AI Model: What’s New?

AI chatbots like Meta AI, ChatGPT, Gemini, etc. are powered by LLMs trained on large data sets. However, these AI models are not readily accessible to the public and software developers as they contain several engineering and trade secrets. Meta has made its Llama 3.1 AI model open-source, allowing developers to implement and deploy it anywhere, without any restrictions.

Llama 3.1 has advanced reasoning skills to solve complex queries like mathematical equations, coding, debugging, and other reasoning-related tasks. It can also provide step-by-step explanations while generating a response. The new AI model also introduces improved support for image generation, allowing users to create avatars by analyzing their faces.

Coming to the developer side of things, Meta’s open-source Llama 3.1 model is faster and more efficient than OpenAI’s GPT-4, GPT-4o, Claude 3.5 Sonnet, and many other flagship-grade AI models. Meta’s model is generating higher scores in IFEval, GSMBK, and Arc Challenge, which test the reasoning capabilities of the AI model.

meta ai llama 3.1
Comparison of Llama 3.1 with other AI models

The biggest advantage of the Llama 3.1 model is its reduced implementation costs. It is said to cost just half as much as GPT-4o to run in production. To recall, GPT-4o was already designed to be a cheaper model by OpenAI, and Meta’s Llama 3.1 model now undercuts it at an even lower price.

Software developers can now implement Llama 3.1 across their projects. Meta CEO Mark Zuckerberg also mentioned that the company is committed to making AI an open-source environment to ensure maximum benefits and safety for all users.

The new Llama 3.1 model has started rolling out to all users in the Meta AI chatbot. It can be accessed on the Meta AI website, WhatsApp, Facebook, Instagram, and Messenger apps. Meta AI continues to be available for free to all users. The chatbot has also introduced support for new languages, which include Hindi, Hindi-Romanized script (also known as Hinglish), French, German, Spanish, Italian, and many others.

The post Meta Unveils Llama 3.1, World’s Largest Open-Source AI Model appeared first on MySmartPrice.


Viewing all articles
Browse latest Browse all 306

Trending Articles