← All notes

Summary

LLaMA is a collection of foundation language models ranging from 7B to 65B parameters, trained on publicly available datasets. In particular, LLaMA 13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA 65B is competitive with the best models, Chinchilla 70B and PaLM 540B. They released the models to the research community. [S1]

Claims

• id: C1 claim: 'LLaMA is a collection of models ranging from 7B to 65B parameters.' source: S1 evidence: "foundation language models ranging from 7B to 65B parameters"
• id: C2 claim: 'They report that LLaMA-13B outperforms GPT-3 (175B) on most benchmarks.' source: S1 evidence: "LLaMA-13B outperforms GPT-3 (175B) on most benchmarks"
• id: C3 claim: 'They released the models to the research community.' source: S1 evidence: "We release all our models to the research community"

Verification

• Code available: Paper: https://github.com/facebookresearch/llama, which returned HTTP 200 and redirected to https://github.com/meta-llama/llama on 2026-11-11. Not run.
• Reproduced: not reproduced (2026-11-11). Code not run.
• Related work: Live notes 2401-04088-mixtral-of-experts and 2210-11416-flan-instruction-finetuning. This is the paper-verification note for LLaMA itself.
• Limits: No dedicated Limitations section was found in the HTML. The paper presents pretrained base models trained on public data and does not itself report instruction-tuning results.
• License: CC BY 4.0 on the abs page (2026-11-11; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 27 Feb 2023 (v1). Sources: [S1] https://arxiv.org/abs/2302.13971, v1 (27 Feb 2023). checked_on: 2026-11-11.

When to apply

Use when comparing early openly released foundation models, or when tracing later Llama-family work back to the original 7B-to-65B release.

Facts

• id: F1 claim: 'LLaMA is a collection of foundation language models ranging from 7B to 65B parameters.' confidence: medium date: 2026-11-11 sources: [S1] license: CC BY 4.0
• id: F2 claim: 'The authors report that LLaMA-13B outperforms GPT-3 (175B) on most benchmarks.' confidence: medium date: 2026-11-11 sources: [S1] license: CC BY 4.0
• id: F3 claim: 'They report that LLaMA-65B is competitive with Chinchilla-70B and PaLM-540B.' confidence: medium date: 2026-11-11 sources: [S1] license: CC BY 4.0

Sources

• [S1] arXiv 2302.13971, https://arxiv.org/abs/2302.13971. License: CC BY 4.0. v1 (27 Feb 2023). Accessed 2026-11-11.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.

@misc{iil_note_bw6srg8w,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{LLaMA: Open and Efficient Foundation Language Models}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2302-13971-llama/},
  note = {Permanent ID iil:note:bw6srg8w. 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/2302-13971-llama.md.