As per the latest news from Cointelegraph, David Sacks, the White House AI and crypto advisor, has declared that the United States has taken the front seat in the global artificial intelligence (AI) race, thanks to Meta’s recent release of its Llama 4 AI model.
“To secure our position in the AI race, it’s crucial for the US to make significant strides in open source, and Llama 4 has allowed us to regain our leading position,” expressed Sacks in an X post dated April 5. This comes amid the increasing speculations surrounding the fierce competition for the top spot in the global AI race between the USA and China.
Having been vocal about the AI race since assuming his role after former US President Donald Trump’s inauguration on January 20, Sacks stated with certainty that he is “confident about the US’s prospects, but complacency is not an option.”
Meta’s AI division disclosed in an X post on the same day that they are rolling out the fourth iteration of their Llama models, Llama 4 Scout and Llama 4 Maverick. As per Meta, these models are “our most sophisticated models to date and lead their class in multimodality.”
The Llama 4 Scout model by Meta boasts of 17 billion active parameters and utilizes 16 experts. Meta asserts that it surpasses its competitors — Gemma 3, Gemini 2.0 Flash-lite, and Mistral 3.1 — “across a wide array of widely recognized benchmarks.” The Llama 4 Maverick model also possesses 17 billion active parameters but is equipped with 128 experts. It is claimed to outperform GPT-4o and Gemini 2.0 Flash “across a wide array of widely recognized benchmarks.”
In July 2024, Meta’s CEO Mark Zuckerberg predicted that by 2025, Llama models would evolve to be “the most advanced in the sector.” Since the release of the limited version of Llama 1 in February 2023, Meta has been overwhelmed by the demand, receiving over 100,000 requests for access.
