{
  "generated": "2026-09-08",
  "benchmark": "visualwebarena",
  "results": [
    {
      "system": "GPT-4V (SoM)",
      "developer": "OpenAI",
      "value": 16.37,
      "date": "2024-01-24",
      "source": {
        "url": "https://arxiv.org/abs/2401.13649",
        "kind": "paper",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "scaffold": "VisualWebArena SoM agent",
        "notes": "Paper baseline, Set-of-Marks + captions + image inputs."
      }
    },
    {
      "system": "GPT-4o (SoM)",
      "developer": "OpenAI",
      "value": 19.78,
      "date": "2024-06-01",
      "source": {
        "url": "https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?gid=2044883967#gid=2044883967",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "scaffold": "VisualWebArena SoM agent",
        "notes": "Maintainer sheet, month only; 1st used."
      }
    },
    {
      "system": "GPT-4o + R-MCTS (ExACT)",
      "developer": "ExACT (leaderboard result source)",
      "value": 33.7,
      "date": "2024-10-01",
      "source": {
        "url": "https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?gid=2044883967#gid=2044883967",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "scaffold": "ExACT R-MCTS",
        "notes": "Month only; 1st used."
      }
    },
    {
      "system": "Gemini 2.5 Flash (SGV)",
      "developer": "Google DeepMind",
      "value": 54.0,
      "date": "2025-07-01",
      "source": {
        "url": "https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?gid=2044883967#gid=2044883967",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "scaffold": "SGV",
        "notes": "Top of the sheet as of access date. Month only; 1st used."
      }
    },
    {
      "system": "GPT-5 (WALT)",
      "developer": "OpenAI",
      "value": 52.9,
      "date": "2025-10-01",
      "source": {
        "url": "https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?gid=2044883967#gid=2044883967",
        "kind": "official-leaderboard",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "scaffold": "WALT",
        "notes": "Month only; 1st used."
      }
    }
  ]
}