Meta has announced that its new artificial intelligence model is the first open-source system that can compete with major players like OpenAI and Anthropic.
The company revealed in a blog post that its latest model, named “Llama 3.1 405B,” is able to perform well in various tasks compared to its competitors. This advancement could potentially make one of the most powerful AI models accessible without any intermediaries controlling access or usage.
Meta stated, “Developers have the freedom to customize the models according to their requirements, train them on new data sets, and fine-tune them further. This empowers developers worldwide to harness the capabilities of generative AI without sharing any data with Meta, and run their applications in any environment.”
Users of Llama on Meta’s app in the US will benefit from an additional layer of security, as the system is open-source and cannot be mandated for use by other companies.
Meta co-founder Mark Zuckerberg emphasized the importance of open source for the future of AI, highlighting its potential to enhance productivity, creativity, and quality of life while ensuring technology is deployed safely and evenly across society.
While Meta’s model matches the size of competing systems, its true effectiveness will be determined through fair testing against other models like GPT-4o.
Currently, Llama 3.1 405B is only accessible to users in 22 countries, excluding the EU. However, it is expected that the open-source system will expand to other regions soon.
Source: www.theguardian.com
