Meta Uncovers Llama 3: A New Standard in Open-Source AI Models
Meta is the tech giant. It has just pronounced the release of its latest open-source generative AI model: Llama 3. This new development is set to transform the AI industry as Meta claims that Llama 3 sets a new standard among open models available today.
The Llama 3 family includes two models that have been trained on two custom-built 24,000 GPU clusters. This massive computational power has allowed the models to learn and adjust at an extraordinary rate, making them highly effective and operative.
Meta backs its claim by pointing to the Llama 3 models’ scores on popular AI benchmarks like MMLU, ARC, and DROP. Despite constant discussions about their rationality, these standards remain one of the few consistent ways by which AI players like Meta assess their models.
In a head-to-head comparison, Llama 3 8B outclasses other open-source models like Mistral’s and Google’s on at least nine levels. This is a major achievement as it validates the superior performance of Llama 3 over its competitors.
Additionally, the larger-parameter-count Llama 3 model that is Llama 3 70B, is competitive with flagship generative AI models including Gemini 1.5 Pro. This shows that Llama 3 is not only superior among open-source models but also holds its own against exclusive AI models.
To confirm the performance of Llama 3, Meta developed its own test set covering use cases ranging from coding and creating writing to reasoning to summarization. In these tests, Llama 3 70B came out on top against Mistral’s Mistral Medium model, OpenAI’s GPT-3.5 and Claude Sonnet.
This release of Llama 3 by Meta marks a breakthrough in the development of open-source AI models. It not only sets a new standard for open models but also defies proprietary models in terms of performance and efficiency. As AI continues to progress, it will be interesting to see how other players in the industry respond to this development.
Read More: CommentSold Modernizes E-commerce with AI-Powered Shoppable Clips
Read More: OpenAI Reveals Enhanced GPT-4 Turbo, Making ChatGPT More Direct and Less Verbal
