Llama 3.2 (1B, 3B)

Training data disclosure: Vague

Level determined for 13% of benchmark cells.

• TestTest data seen0
• SeenSeen, split unclear0
• TrainTraining data seen0
• Rel.Related data seen0
• NoNot seen2
• ?Unknown13

Other declared training data

• A new mix of publicly available online data. (broad web data: may contain benchmark test sets) · split: unknown
  Evidence

  “A new mix of publicly available online data.”

  “Llama 3.2 was pretrained on up to 9 trillion tokens of data from publicly available sources.”

• logits from Llama 3.1 8B and 70B (knowledge distillation) (not yet matched to a dataset) · split: unknown

  Teacher-model outputs, not a dataset; see llama-3.1 for the teachers' contamination findings.

  Evidence

  “For the 1B and 3B Llama 3.2 models, we incorporated logits from the Llama 3.1 8B and 70B models into the pretraining stage of the model development, where outputs (logits) from these larger models were used as token-level targets.”

• human-generated data from vendors combined with synthetic data (not yet matched to a dataset) · split: unknown · fine-tuning
  Evidence

  “We employ a multi-faceted approach to data collection, combining human-generated data from our vendors with synthetic data to mitigate potential safety risks.”

  “In post-training we used a similar recipe as Llama 3.1 and produced final chat models by doing several rounds of alignment on top of the pre-trained model.”

Benchmark cells

Classic LLM benchmarks

Recent LLM benchmarks

Notes

1B/3B models were obtained with pruning and knowledge distillation from Llama 3.1 8B/70B (model card). No contamination analysis published for this release.

Sources

• Llama 3.2 MODEL_CARD.md (meta-llama/llama-models) · model card · retrieved 2026-10-15
• meta-llama/llama-models README (launch dates) · readme · retrieved 2026-10-15