Meta launches ‘most capable openly available LLM to date,’ rivaling GPT and Claude

Meta raised the artificial intelligence (AI) bar on April 18 by announcing that its newest large language model, Llama-3, is the “most capable” and “best open source model” currently available. 

The company’s statements surrounding the general availability of Llama-3, as well as a new standalone “Meta AI” portal, has the tech world abuzz with declarations that the current space leaders — Microsoft, OpenAI, Google and Anthropic — finally have some stiff competition from the company formerly called Facebook.

“This next generation of Llama demonstrates state-of-the-art performance on a wide range of industry benchmarks and offers new capabilities, including improved reasoning. We believe these are the best open source models of their class, period.”

The large language model market continues to expand year over year, with updates to the most popular models seemingly occurring at a steady clip. Current market leader OpenAI is rumored to be on the verge of launching its latest model — GPT-5, if it follows previous naming conventions.

While the costs continue to rise, the rewards could be even greater. At stake, according to both Google and Microsoft, is the development of the world’s first artificial general intelligence (AGI). An AGI system, hypothetically speaking, would be capable of doing any task a human could do, given the proper resources.