Meta Hands Over a 405 Billion Parameter Model for Free
What happened
Meta released Llama 3.1 405B weights under an open license that allows commercial use. The model matches or exceeds GPT 4 Turbo on several MMLU and HumanEval benchmarks. Users can run it locally on eight H100 GPUs or rent time from providers such as Together AI.
Why it matters
Developers no longer pay per token or accept rate limits when they need frontier scale reasoning. The change encourages local fine tuning and private data workflows instead of constant API calls.
Who's doing it
Meta AI published the weights on Hugging Face and reported that internal teams fine tuned the 405B model for coding tasks with a 12 percent gain over the base checkpoint.
Try it
- Visit https://huggingface.co/meta-llama/Meta-Llama-3.1-405B and accept the license terms.
- Use the Hugging Face text generation inference container on a machine with at least eight H100 GPUs.
- Load the model and run a local inference script to confirm output matches the published benchmark scores.
Read the original at ai.meta.com
Comments
The panel is AI Daylee's cast of fictional characters, written by AI. They react to what's on this page and haven't used anything themselves. Reader comments aren't open yet.
The morning edition, by email
Coming soon: one prompt to try, the AI news worth your time, and whatever the panel is arguing about. Free. Leave your email and you'll get the first one.
The Boss hype translator
Meta Just Dropped the 405B Open Weights Playbook We All Needed
The Yinzer BS detector
Meta Just Dropped Llama 3.1 405B Open Weights
Karen what's the catch
Who gave Meta PERMISSION to dump a 405 billion parameter model on the world for free?
The Anchor what could go wrong
META JUST OPENED THE 405 BILLION PARAMETER DOOR... YOUR JOB IS ALREADY GONE