You can now try Meta's new Llama 3.1 405B on Amazon Bedrock & Azure AI
It's been an eventful week at Meta.
2 min. read
Published on
Share this article
Improve this guide
Read our disclosure page to find out how can you help MSPoweruser sustain the editorial team Read more
Key notes
- Meta just launched the Meta Llama 3.1 405B model, its most advanced to this date.
- It excels in tasks such as synthetic data generation, long-form text generation, multilingual translation, etc.
- The model is now available across various platforms.
It’s been quite an eventful week at Meta. The Facebook parent company has recently launched a new feature on Meta AI called “Imagine Me,” which lets you envision yourself anywhere and in anyone. Besides, the tech giant also launched its latest AI model, Llama 3.1 405B, and it’s available now across different platforms.
Meta boasts that the model, which is currently in preview, is currently the largest publicly available large language model (LLM) from the tech giant. It also comes with a 128K context length and improved multilingual for the 8B and 70B variants. Here’s how it’s compared to OpenAI’s GPT-4 family:
The tech giant also says that the model excels in tasks such as synthetic data generation, long-form text generation, multilingual translation, coding, math, tool use, enhanced contextual understanding, and advanced reasoning and decision-making.
Microsoft, in collaboration with Meta, also announced the availability of the model on Azure AI through a serverless API endpoint. This model, along with fine-tuned versions of Llama 3.1 8B and 70B, can be accessed via Azure AI Studio.
It wasn’t too long ago that Meta launched the Llama-3 400B model. At its launch, the model showed promise, scoring competitively in benchmarks such as MMLU, GPQA, HumanEval, and MATH, though slightly behind established models like OpenAI’s GPT-4 Turbo and Anthropic’s Claude 3 Opus.
Meta’s Llama 3.1 405B is also available on other platforms like Amazon Bedrock, Databricks’ Mosaic AI Playground, and others.
