Meta released LLaMA and, after a leak, catalyzed an open fine-tune ecosystem that challenged closed APIs.
In February 2023, Meta presented LLaMA, efficient language models for research. A weight leak accelerated community replicas and fine-tunes (Alpaca, Vicuna).
LLaMA 2 (July 2023) arrived with a more permissive commercial license. The message: not every frontier will sit behind an API paywall.
The open/closed duality has since defined AI cost, security, and technological sovereignty.
New Llama generations improved multimodality and agents while keeping downloadable weights as industrial strategy.
Meta, Mistral, DeepSeek, and others brought public weights near closed-API performance, redefining “frontier.”
Meta Superintelligence Labs shipped Muse Spark 1.1 with Thinking mode on meta.ai and a public preview of the Meta Model API, nearly a year after Llama 4.