Meta Hands Over a 405-Billion-Parameter Model for Free
Meta published Llama 3.1 405B under an open license. Anyone can download the full weights and run inference locally or on rented GPUs. No per-token charges apply.
This shifts the cost structure from usage fees to fixed hardware. Teams can now prototype and deploy large models without metering anxiety. The change rewards experimentation over rationing prompts.
Hugging Face hosts the weights and reports over 100,000 downloads in the first week. Startups are already fine-tuning the model on commodity cloud instances.
Step 1: Go to huggingface.co/meta-llama and accept the license terms. Step 2: Download the 405B weights or load them via the Transformers library. Step 3: Run inference on a rented A100 or H100 instance and compare output quality to paid frontier APIs.