Meta released Llama 4, the latest version of its open-source large language model family, for free download under a commercial licence — a direct challenge to the closed, subscription-based AI ecosystems of OpenAI, Google and Anthropic.
Llama 4 is available in three sizes — 8B, 70B and 405B parameters — with the largest version matching or exceeding GPT-4o performance on eight of twelve standard benchmarks. The model supports 32 languages and a 128,000-token context window.
Comments (3)
Leave a comment
The headline doesn't do the article justice. The real story is buried in paragraph six.
This deserves a much wider audience than it'll probably get. Mainstream media ignores stories like this.
As someone who has worked in this sector for 15 years, I can say the reporting here is accurate and the implications are as serious as the article suggests. The question is whether policymakers will act before it's too late.
Good journalism. Rare these days. Thank you.