{
  "title": "AI progress in the 2023 report",
  "report_year": 2023,
  "period": "Historical snapshot: October 2023. Dates and populations are specified per row.",
  "methodology": "Model size and corpus size are specifications, not quality scores. Comparisons in the report used different tasks and evaluators. Llama’s availability was subject to its license; small-model findings were limited to the evaluated settings.",
  "source_deck": "https://docs.google.com/presentation/d/156WpBF_rGvf4Ecg19oM1fyR51g4FAmHV3Zs0WLukrLQ/edit",
  "web_edition_date": "2026-10-10",
  "rows": [
    {
      "measure": "Llama 2 pretraining corpus",
      "value": "2T tokens",
      "definition": "Corpus size reported for Llama 2, a 40% increase over its predecessor.",
      "printed_slide": 19,
      "pdf_page": 19,
      "source_url": "https://www.stateof.ai/State-of-AI-Report-2023.pdf#page=19"
    },
    {
      "measure": "Llama 2 largest model discussed",
      "value": "70B parameters",
      "definition": "Model compared with ChatGPT across tasks in the report.",
      "printed_slide": 19,
      "pdf_page": 19,
      "source_url": "https://www.stateof.ai/State-of-AI-Report-2023.pdf#page=19"
    },
    {
      "measure": "phi-1 model size",
      "value": "1.3B parameters",
      "definition": "Small code model trained on curated code and synthetic educational material.",
      "printed_slide": 26,
      "pdf_page": 26,
      "source_url": "https://www.stateof.ai/State-of-AI-Report-2023.pdf#page=26"
    }
  ]
}