The news, 365 days behind — on purpose Delayed live · replaying 2025

One Year Ago.AI

Remember how fast this is.

23JUL2024replayed
one year on
model launchMeta · OpenAI · Anthropic

Meta Releases Llama 3.1 405B, Says It Rivals Leading Closed Models

The open-source model adds 128K context to the 8B and 70B versions and updates the license to let developers use outputs for synthetic data generation and model distillation.

Meta today released Llama 3.1 405B, a 405-billion-parameter model it says is the first frontier-level open source AI model to rival leading proprietary systems like OpenAI’s GPT-4o and Anthropic’s Claude 3.5 Sonnet. Trained on over 15 trillion tokens using 16,000 Nvidia H100 GPUs, the model is text-only but achieves competitive results across coding, math, multilingual translation, and general knowledge benchmarks.

Alongside the flagship, Meta refreshed its 8B and 70B models with 128,000-token context windows — a 16x increase over the prior Llama 3 generation — and added support for eight languages and tool use. The company also updated its license to explicitly allow developers to use outputs from Llama models to train other models, which Meta says helps with synthetic data generation and model distillation, addressing a major criticism from the open-source community. However, the license still restricts developers with over 700 million monthly active users, who must seek special permission.

In a simultaneous open letter, CEO Mark Zuckerberg argues that open source is good for developers, Meta and the world, and frames the release as part of Meta’s open-access push. The release comes with a reference system, safety tools (Llama Guard 3, Prompt Guard), and a call for comment on a proposed Llama Stack API. More than 25 launch partners — including AWS, Nvidia, Databricks, Groq, and Google Cloud — announced support on day one.

Critics note that the model still exhibits hallucinations, requires a server node for inference, and trails Claude 3.5 Sonnet on some reasoning tasks. Meta says the model can be downloaded, fine-tuned and deployed without sharing data with Meta.

One year later — open only if you can handle spoilers

Llama 3.1 405B became a go-to base model for the open-source community, but its 16,000-GPU requirement limited widespread local use; it was eventually spun off into smaller distilled versions that achieved broader adoption. The debate over the license's user cap did not trigger major forks, but several startup lobbied Meta to relax the threshold.

Replay thisPost on XRedditHNLinkedIn

The Weekly Replay · free by email

This week, one year ago — every Sunday.

One email each Sunday: the week's replayed AI news, with the one-year-later annotations included. Written like it's breaking — dated like it isn't.

Free · double opt-in · unsubscribe anytime · privacy