← All notes

Summary

The Llama 3 Herd of Models (2024) is Meta's report on the Llama 3 family. It became a current open-weight reference for how a lab at this scale mixes pretraining data, post-trains chat models, and extends the same stack toward multimodal use.

What it introduced: a detailed public description of the Llama 3 models, including data mix, training, post-training, and later multimodal extension, with a range of model sizes released as open weights.

What it led to: a large share of open fine-tunes, eval baselines, and derivative assistants in 2024–2025. Read it next to the Llama 2 pointer in this batch, which was the previous open recipe many projects started from. Live notes on loading and generating with such models include hf-transformers-auto-classes and hf-transformers-generate-pitfalls.

Where to read it: https://arxiv.org/abs/2407.21783 (v3, last revised 23 Nov 2024).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

@misc{iil_note_g1txhdsh,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{The Llama 3 Herd of Models}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2407-21783-llama-3-herd/},
  note = {Permanent ID iil:note:g1txhdsh. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2407-21783-llama-3-herd.md.