Meta recently launched Llama 3, the latest iteration in its series of large language models. The first version of the Llama model was released in February of last year as one of the first open-weight large language models. Subsequently, it launched the second version in July 2023.
After Llama 2 gained widespread acceptance and popularity among researchers and developers, Llama 3 overpowered other existing open models, such as Mistral and Gemma, in the majority of the performance benchmarks. Scroll down to know and understand more about Meta AI Open LLM Performance, Benchmarks, Price, and Other Details.
What is Llama 3?
Llama 3 is a text-generation AI model, similar to OpenAI’s GPT and Anthropic’s Claude model. It is an accessible, open-source large language model (LLM) designed for developers, researchers, and businesses to build, experiment with, and responsibly scale their generative AI ideas.
With enhanced scalability and performance, Llama 3 can handle multi-step tasks effortlessly, while our refined post-training processes significantly lower false refusal rates, improve response alignment, and boost diversity in model answers. Additionally, it drastically elevates capabilities like reasoning, code generation, and instruction following.
According to Hugging Face, this smart assistant comes in two sizes: 8B for efficient deployment and development on consumer-size GPU, and 70B for large-scale AI native applications. Both come in base and instruction-tuned variants.
Rabbit R1 vs Humane AI Pin vs. Limitless Pendant: Which AI Wearable Device is Better?
How is Llama 3 different from Llama 2?
The adoption of a new tokenizer in Llama 3 compared to Llama 2 is a significant change, which increases the vocabulary size to 128,256 (from 32K tokens in the previous version). This larger vocabulary can encode text more efficiently (both for input and output) and potentially yield stronger multilingualism. However, this comes at a price: the small model’s parameter count increases from 7B in Llama 2 to 8B in Llama 3, largely due to the larger embedding input and output matrices. Also, the 8B version of the model now uses Grouped-Query Attention (GQA), which is an efficient representation that should help with longer contexts.
How To Get Rid Of My AI On Snapchat? Easy Steps To Follow
What are the benchmarks of Llama 3?
Llama 3 models take data and scale to new heights. It has been trained on two newly disclosed custom-built 24K GPU clusters with over 15T tokens of data; this represents a training dataset that is 7 times larger and contains 4 times more code than that used for Llama 2. This results in the most capable Llama model yet, which supports an 8K context length that doubles the capacity of Llama 2.
As per the Meta blog, this evaluation set contains 1,800 prompts that cover 12 key use cases: asking for advice, brainstorming, classification, closed question answering, coding, creative writing, extraction, inhabiting a character/persona, open question answering, reasoning, rewriting, and summarization.
The chart below shows the aggregated results of our human evaluations across different categories and prompts against Claude Sonnet, Mistral Medium, and GPT-3.5.
Preference rankings by human annotators based on this evaluation set highlight the strong performance of the 70B instruction-following model compared to competing models of comparable size in real-world scenarios.
The pre-trained model also establishes a new state-of-the-art for LLM models at those scales.
What is the price of Meta Smart Assistant?
At present, the price for Meta Llama 3 is not available in the public domain.
You can use Meta AI on Facebook, Instagram, WhatsApp, Messenger, and the web to get things done, learn, create, and connect with the things that matter to you. Also, you can find all 5 open-access models (2 base models, 2 fine-tuned and the Llama Guard) on the Hugging Face.
Meta further aims for the latest model to be multilingual and multimodal, have a longer context, and continue to improve overall performance across core LLM capabilities such as reasoning and coding. Visit the Llama 3 website to download the models and reference the Getting Started Guide for the latest list of all available platforms.
