Meta Unleashes Llama 3.1 405B. The Weights Are Open. The Compute Bill Is Not.
Meta released Llama 3.1 405B as open weights with a commercial license, making it the largest open-source model ever released. The model can be downloaded and run locally or deployed commercially without restrictions from Meta. Small businesses can build AI products without ongoing API costs or usage limits imposed by big tech companies.
This illustrates the concept of weight liberation, where model parameters shift from gated service to transferable artifact. The mechanism is open-weight licensing with commercial permission, which separates the cost of model access from the cost of compute infrastructure. The lesson is that ownership of the model and ownership of the hardware to run it are now distinct economic decisions, and the latter remains the real bottleneck.
Meta released Llama 3.1 405B under a commercial license at ai.meta.com/blog/meta-llama-3-1. The model is downloadable and deployable without API costs or usage restrictions, which Meta positions as the largest open-source model release to date.
- Visit ai.meta.com/blog/meta-llama-3-1 and review the download instructions and license terms. Confirm your intended use qualifies under the commercial license before proceeding.
- Since running a 405 billion parameter model locally requires hardware most consumers lack, instead download a smaller Llama variant through a consumer platform like Hugging Face or Ollama. Install Ollama and run a command like 'ollama run llama3.1' to experience a locally hosted open-weight model.
- Compare responses from your local Llama model against a hosted service like ChatGPT. Note the differences in speed, quality, and the fact that your local instance has no per-query cost and no data leaving your machine.