LLaMA

Research only · Language · Milestone

LLaMA is a set of large language models for text from Meta, released on 24 February 2023 in sizes of 7B, 13B, 33B and 65B parameters to help researchers study such models. The weights were shared case by case with approved researchers under a non-commercial licence; from the Llama 2 launch on, Meta calls it Llama 1. [1] [2]README: Download [3] Primary source

Timeline of LLaMA →

Claims and evidence

  • Released Primary source[1]page date: February 24, 2023
  • Status Research onlyWeights are sent only after an approved request, under a non-commercial licence for research use. The request form linked from the llama_v1 README was still online on 2026-10-01; no Meta page says whether requests are still granted. Meta's Hugging Face organisations hold no LLaMA repos. Primary source[1]paragraph on the noncommercial license [2]README: Download [4]Model details: License; Intended use
      • Input text Primary source[1]paragraph beginning Like other large language models
      • Output text Primary source[1]paragraph beginning Like other large language models
      • Context window 2k tokensStated in the Llama 2 paper's comparison of Llama 1 and Llama 2; the LLaMA paper and model card give no figure. Primary source[5]Table 1
      • Access research only Primary source[1]paragraph on the noncommercial license [6]Abstract [2]README: Download

      Lineage

      Predecessors

      No known predecessor.

      Successors

      Based on

      Not derived from another model.

      Variants and derived

      • IDEFICS · Hugging Face · 22 August 2023 · derived (other)

      Siblings

      None recorded.

      All ancestors

      None.

      All descendants

      Variants

      LLaMA 7B

      Same dates as LLaMA.

      • Variant Inline variant in this record. Primary source[1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      Also known as: LLaMA-7B

      Differs in:

      • Parameters: 7B Evidence not assessed [1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      LLaMA 13B

      Same dates as LLaMA.

      • Variant Inline variant in this record. Primary source[1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      Also known as: LLaMA-13B

      Differs in:

      • Parameters: 13B Evidence not assessed [1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      LLaMA 33B

      Same dates as LLaMA.

      • Variant Inline variant in this record.The download script on the llama_v1 branch names this size 30B. Primary source[1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type [6]Section 2 Approach, Figure 1 caption [2]download.sh

      Also known as: LLaMA-33B, LLaMA 30B, LLaMA-30B

      Differs in:

      • Parameters: 33B Evidence not assessed [1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      LLaMA 65B

      Same dates as LLaMA.

      • Variant Inline variant in this record. Primary source[1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      Also known as: LLaMA-65B

      Differs in:

      • Parameters: 65B Evidence not assessed [1]paragraph listing the 7B, 13B, 33B and 65B sizes [4]Model details: Model type

      Related AI Radar coverage

      AI Radar coverage starts in June 2026; no coverage linked yet.

      All model releases from Meta on AI Radar →