[
    {
        "id": "77f25849-862250c4",
        "method": {
            "name": "PlanExec",
            "tooltip": "Plan & Execute"
        },
        "llm": {
            "name": "LLaMA3-70B",
            "tooltip": "meta-llama/Llama-3-70b-chat-hf"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 8.9,
                "scenario_goal_completion": 1.8,
                "interactions": 33.72
            },
            "level 1": {
                "task_goal_completion": 19.3,
                "scenario_goal_completion": 5.3,
                "interactions": 33.72
            },
            "level 2": {
                "task_goal_completion": 8.3,
                "scenario_goal_completion": 0.0,
                "interactions": 33.72
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 33.72
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 2.4,
                "scenario_goal_completion": 0.7,
                "interactions": 63.09
            },
            "level 1": {
                "task_goal_completion": 11.1,
                "scenario_goal_completion": 4.2,
                "interactions": 63.09
            },
            "level 2": {
                "task_goal_completion": 0.7,
                "scenario_goal_completion": 0.0,
                "interactions": 63.09
            },
            "level 3": {
                "task_goal_completion": 0.5,
                "scenario_goal_completion": 0.0,
                "interactions": 63.09
            }
        }
    },
    {
        "id": "4077db41-07bc5204",
        "method": {
            "name": "ReAct",
            "tooltip": "Reason + Act"
        },
        "llm": {
            "name": "LLaMA3-70B",
            "tooltip": "meta-llama/Llama-3-70b-chat-hf"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 20.8,
                "scenario_goal_completion": 8.9,
                "interactions": 24.43
            },
            "level 1": {
                "task_goal_completion": 42.1,
                "scenario_goal_completion": 15.8,
                "interactions": 24.43
            },
            "level 2": {
                "task_goal_completion": 12.5,
                "scenario_goal_completion": 6.2,
                "interactions": 24.43
            },
            "level 3": {
                "task_goal_completion": 7.9,
                "scenario_goal_completion": 4.8,
                "interactions": 24.43
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 3.4,
                "scenario_goal_completion": 0.0,
                "interactions": 36.0
            },
            "level 1": {
                "task_goal_completion": 12.5,
                "scenario_goal_completion": 0.0,
                "interactions": 36.0
            },
            "level 2": {
                "task_goal_completion": 3.3,
                "scenario_goal_completion": 0.0,
                "interactions": 36.0
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 36.0
            }
        }
    },
    {
        "id": "9ae299b8-25c255f5",
        "method": {
            "name": "ReAct",
            "tooltip": "Reason + Act"
        },
        "llm": {
            "name": "DeepSeekCoder",
            "tooltip": "deepseek-ai/deepseek-coder-33b-instruct"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 7.1,
                "scenario_goal_completion": 1.8,
                "interactions": 46.98
            },
            "level 1": {
                "task_goal_completion": 21.1,
                "scenario_goal_completion": 5.3,
                "interactions": 46.98
            },
            "level 2": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 46.98
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 46.98
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 2.9,
                "scenario_goal_completion": 0.7,
                "interactions": 53.78
            },
            "level 1": {
                "task_goal_completion": 16.7,
                "scenario_goal_completion": 4.2,
                "interactions": 53.78
            },
            "level 2": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 53.78
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 53.78
            }
        }
    },
    {
        "id": "871bf9fd-4de66f57",
        "method": {
            "name": "PlanExec",
            "tooltip": "Plan & Execute"
        },
        "llm": {
            "name": "DeepSeekCoder",
            "tooltip": "deepseek-ai/deepseek-coder-33b-instruct"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 1.8,
                "scenario_goal_completion": 0.0,
                "interactions": 71.65
            },
            "level 1": {
                "task_goal_completion": 5.3,
                "scenario_goal_completion": 0.0,
                "interactions": 71.65
            },
            "level 2": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 71.65
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 71.65
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 0.7,
                "scenario_goal_completion": 0.0,
                "interactions": 55.65
            },
            "level 1": {
                "task_goal_completion": 4.2,
                "scenario_goal_completion": 0.0,
                "interactions": 55.65
            },
            "level 2": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 55.65
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 55.65
            }
        }
    },
    {
        "id": "3a889687-923696fe",
        "method": {
            "name": "IPFunCall",
            "tooltip": "Iterative Parallel Function Calling"
        },
        "llm": {
            "name": "GPT-4o",
            "tooltip": "gpt-4o-2024-05-13"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 32.1,
                "scenario_goal_completion": 16.1,
                "interactions": 17.74
            },
            "level 1": {
                "task_goal_completion": 49.1,
                "scenario_goal_completion": 31.6,
                "interactions": 17.74
            },
            "level 2": {
                "task_goal_completion": 20.8,
                "scenario_goal_completion": 6.2,
                "interactions": 17.74
            },
            "level 3": {
                "task_goal_completion": 25.4,
                "scenario_goal_completion": 9.5,
                "interactions": 17.74
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 18.0,
                "scenario_goal_completion": 10.1,
                "interactions": 16.12
            },
            "level 1": {
                "task_goal_completion": 27.8,
                "scenario_goal_completion": 12.5,
                "interactions": 16.12
            },
            "level 2": {
                "task_goal_completion": 24.0,
                "scenario_goal_completion": 16.0,
                "interactions": 16.12
            },
            "level 3": {
                "task_goal_completion": 9.7,
                "scenario_goal_completion": 4.6,
                "interactions": 16.12
            }
        }
    },
    {
        "id": "c722a232-74b9384b",
        "method": {
            "name": "PlanExec",
            "tooltip": "Plan & Execute"
        },
        "llm": {
            "name": "GPT-4 Turbo",
            "tooltip": "gpt-4-turbo-2024-04-09"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 32.7,
                "scenario_goal_completion": 16.1,
                "interactions": 20.14
            },
            "level 1": {
                "task_goal_completion": 64.9,
                "scenario_goal_completion": 31.6,
                "interactions": 20.14
            },
            "level 2": {
                "task_goal_completion": 27.1,
                "scenario_goal_completion": 18.8,
                "interactions": 20.14
            },
            "level 3": {
                "task_goal_completion": 7.9,
                "scenario_goal_completion": 0.0,
                "interactions": 20.14
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 11.0,
                "scenario_goal_completion": 3.6,
                "interactions": 23.97
            },
            "level 1": {
                "task_goal_completion": 36.1,
                "scenario_goal_completion": 16.7,
                "interactions": 23.97
            },
            "level 2": {
                "task_goal_completion": 10.7,
                "scenario_goal_completion": 2.0,
                "interactions": 23.97
            },
            "level 3": {
                "task_goal_completion": 2.1,
                "scenario_goal_completion": 0.0,
                "interactions": 23.97
            }
        }
    },
    {
        "id": "08d2ff3e-a0589430",
        "method": {
            "name": "IPFunCall",
            "tooltip": "Iterative Parallel Function Calling"
        },
        "llm": {
            "name": "GPT-4 Turbo",
            "tooltip": "gpt-4-turbo-2024-04-09"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 30.4,
                "scenario_goal_completion": 21.4,
                "interactions": 13.02
            },
            "level 1": {
                "task_goal_completion": 43.9,
                "scenario_goal_completion": 31.6,
                "interactions": 13.02
            },
            "level 2": {
                "task_goal_completion": 22.9,
                "scenario_goal_completion": 18.8,
                "interactions": 13.02
            },
            "level 3": {
                "task_goal_completion": 23.8,
                "scenario_goal_completion": 14.3,
                "interactions": 13.02
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 14.6,
                "scenario_goal_completion": 9.3,
                "interactions": 13.56
            },
            "level 1": {
                "task_goal_completion": 22.2,
                "scenario_goal_completion": 16.7,
                "interactions": 13.56
            },
            "level 2": {
                "task_goal_completion": 19.3,
                "scenario_goal_completion": 14.0,
                "interactions": 13.56
            },
            "level 3": {
                "task_goal_completion": 8.2,
                "scenario_goal_completion": 3.1,
                "interactions": 13.56
            }
        }
    },
    {
        "id": "d035381e-2d30c0e3",
        "method": {
            "name": "FullCodeRefl",
            "tooltip": "Full Code + Reflection"
        },
        "llm": {
            "name": "GPT-4o",
            "tooltip": "gpt-4o-2024-05-13"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 33.9,
                "scenario_goal_completion": 26.8,
                "interactions": 3.31
            },
            "level 1": {
                "task_goal_completion": 61.4,
                "scenario_goal_completion": 52.6,
                "interactions": 3.31
            },
            "level 2": {
                "task_goal_completion": 29.2,
                "scenario_goal_completion": 25.0,
                "interactions": 3.31
            },
            "level 3": {
                "task_goal_completion": 12.7,
                "scenario_goal_completion": 4.8,
                "interactions": 3.31
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 19.2,
                "scenario_goal_completion": 12.2,
                "interactions": 4.39
            },
            "level 1": {
                "task_goal_completion": 54.2,
                "scenario_goal_completion": 41.7,
                "interactions": 4.39
            },
            "level 2": {
                "task_goal_completion": 18.7,
                "scenario_goal_completion": 10.0,
                "interactions": 4.39
            },
            "level 3": {
                "task_goal_completion": 6.7,
                "scenario_goal_completion": 3.1,
                "interactions": 4.39
            }
        }
    },
    {
        "id": "ac5b742b-aa0243da",
        "method": {
            "name": "PlanExec",
            "tooltip": "Plan & Execute"
        },
        "llm": {
            "name": "GPT-4o",
            "tooltip": "gpt-4o-2024-05-13"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 44.6,
                "scenario_goal_completion": 23.2,
                "interactions": 31.62
            },
            "level 1": {
                "task_goal_completion": 75.4,
                "scenario_goal_completion": 47.4,
                "interactions": 31.62
            },
            "level 2": {
                "task_goal_completion": 39.6,
                "scenario_goal_completion": 18.8,
                "interactions": 31.62
            },
            "level 3": {
                "task_goal_completion": 20.6,
                "scenario_goal_completion": 4.8,
                "interactions": 31.62
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 19.7,
                "scenario_goal_completion": 7.9,
                "interactions": 40.74
            },
            "level 1": {
                "task_goal_completion": 50.0,
                "scenario_goal_completion": 29.2,
                "interactions": 40.74
            },
            "level 2": {
                "task_goal_completion": 17.3,
                "scenario_goal_completion": 6.0,
                "interactions": 40.74
            },
            "level 3": {
                "task_goal_completion": 10.3,
                "scenario_goal_completion": 1.5,
                "interactions": 40.74
            }
        }
    },
    {
        "id": "0465049f-bff9fca2",
        "method": {
            "name": "FullCodeRefl",
            "tooltip": "Full Code + Reflection"
        },
        "llm": {
            "name": "DeepSeekCoder",
            "tooltip": "deepseek-ai/deepseek-coder-33b-instruct"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 13.1,
                "scenario_goal_completion": 8.9,
                "interactions": 4.17
            },
            "level 1": {
                "task_goal_completion": 26.3,
                "scenario_goal_completion": 15.8,
                "interactions": 4.17
            },
            "level 2": {
                "task_goal_completion": 12.5,
                "scenario_goal_completion": 12.5,
                "interactions": 4.17
            },
            "level 3": {
                "task_goal_completion": 1.6,
                "scenario_goal_completion": 0.0,
                "interactions": 4.17
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 5.8,
                "scenario_goal_completion": 2.9,
                "interactions": 5.46
            },
            "level 1": {
                "task_goal_completion": 29.2,
                "scenario_goal_completion": 12.5,
                "interactions": 5.46
            },
            "level 2": {
                "task_goal_completion": 2.0,
                "scenario_goal_completion": 2.0,
                "interactions": 5.46
            },
            "level 3": {
                "task_goal_completion": 0.0,
                "scenario_goal_completion": 0.0,
                "interactions": 5.46
            }
        }
    },
    {
        "id": "baf95530-cee120fc",
        "method": {
            "name": "FullCodeRefl",
            "tooltip": "Full Code + Reflection"
        },
        "llm": {
            "name": "LLaMA3-70B",
            "tooltip": "meta-llama/Llama-3-70b-chat-hf"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 24.4,
                "scenario_goal_completion": 17.9,
                "interactions": 3.75
            },
            "level 1": {
                "task_goal_completion": 50.9,
                "scenario_goal_completion": 36.8,
                "interactions": 3.75
            },
            "level 2": {
                "task_goal_completion": 22.9,
                "scenario_goal_completion": 18.8,
                "interactions": 3.75
            },
            "level 3": {
                "task_goal_completion": 1.6,
                "scenario_goal_completion": 0.0,
                "interactions": 3.75
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 7.0,
                "scenario_goal_completion": 4.3,
                "interactions": 4.19
            },
            "level 1": {
                "task_goal_completion": 34.7,
                "scenario_goal_completion": 20.8,
                "interactions": 4.19
            },
            "level 2": {
                "task_goal_completion": 0.7,
                "scenario_goal_completion": 0.0,
                "interactions": 4.19
            },
            "level 3": {
                "task_goal_completion": 1.5,
                "scenario_goal_completion": 1.5,
                "interactions": 4.19
            }
        }
    },
    {
        "id": "3414a75f-3df6a61f",
        "method": {
            "name": "ReAct",
            "tooltip": "Reason + Act"
        },
        "llm": {
            "name": "GPT-4o",
            "tooltip": "gpt-4o-2024-05-13"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 48.8,
                "scenario_goal_completion": 32.1,
                "interactions": 23.55
            },
            "level 1": {
                "task_goal_completion": 73.7,
                "scenario_goal_completion": 57.9,
                "interactions": 23.55
            },
            "level 2": {
                "task_goal_completion": 52.1,
                "scenario_goal_completion": 31.2,
                "interactions": 23.55
            },
            "level 3": {
                "task_goal_completion": 23.8,
                "scenario_goal_completion": 9.5,
                "interactions": 23.55
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 30.2,
                "scenario_goal_completion": 13.0,
                "interactions": 34.24
            },
            "level 1": {
                "task_goal_completion": 58.3,
                "scenario_goal_completion": 33.3,
                "interactions": 34.24
            },
            "level 2": {
                "task_goal_completion": 28.7,
                "scenario_goal_completion": 12.0,
                "interactions": 34.24
            },
            "level 3": {
                "task_goal_completion": 21.0,
                "scenario_goal_completion": 6.2,
                "interactions": 34.24
            }
        }
    },
    {
        "id": "571aa685-a8671c57",
        "method": {
            "name": "FullCodeRefl",
            "tooltip": "Full Code + Reflection"
        },
        "llm": {
            "name": "GPT-4 Turbo",
            "tooltip": "gpt-4-turbo-2024-04-09"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 25.6,
                "scenario_goal_completion": 19.6,
                "interactions": 3.07
            },
            "level 1": {
                "task_goal_completion": 43.9,
                "scenario_goal_completion": 36.8,
                "interactions": 3.07
            },
            "level 2": {
                "task_goal_completion": 35.4,
                "scenario_goal_completion": 25.0,
                "interactions": 3.07
            },
            "level 3": {
                "task_goal_completion": 1.6,
                "scenario_goal_completion": 0.0,
                "interactions": 3.07
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 12.5,
                "scenario_goal_completion": 7.2,
                "interactions": 4.66
            },
            "level 1": {
                "task_goal_completion": 47.2,
                "scenario_goal_completion": 33.3,
                "interactions": 4.66
            },
            "level 2": {
                "task_goal_completion": 9.3,
                "scenario_goal_completion": 4.0,
                "interactions": 4.66
            },
            "level 3": {
                "task_goal_completion": 2.1,
                "scenario_goal_completion": 0.0,
                "interactions": 4.66
            }
        }
    },
    {
        "id": "1f0055c0-123849bb",
        "method": {
            "name": "ReAct",
            "tooltip": "Reason + Act"
        },
        "llm": {
            "name": "GPT-4 Turbo",
            "tooltip": "gpt-4-turbo-2024-04-09"
        },
        "url": "https://appworld.dev",
        "date": "2024-07-26",
        "test_normal": {
            "all": {
                "task_goal_completion": 26.8,
                "scenario_goal_completion": 12.5,
                "interactions": 18.75
            },
            "level 1": {
                "task_goal_completion": 50.9,
                "scenario_goal_completion": 26.3,
                "interactions": 18.75
            },
            "level 2": {
                "task_goal_completion": 31.2,
                "scenario_goal_completion": 12.5,
                "interactions": 18.75
            },
            "level 3": {
                "task_goal_completion": 1.6,
                "scenario_goal_completion": 0.0,
                "interactions": 18.75
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 17.5,
                "scenario_goal_completion": 5.8,
                "interactions": 26.61
            },
            "level 1": {
                "task_goal_completion": 47.2,
                "scenario_goal_completion": 20.8,
                "interactions": 26.61
            },
            "level 2": {
                "task_goal_completion": 16.7,
                "scenario_goal_completion": 6.0,
                "interactions": 26.61
            },
            "level 3": {
                "task_goal_completion": 7.2,
                "scenario_goal_completion": 0.0,
                "interactions": 26.61
            }
        }
    },
    {
        "id": "e5e08b7e-d8e6474f",
        "method": {
            "name": "LOOP",
            "tooltip": "Leave-one-out PPO"
        },
        "llm": {
            "name": "Qwen2.5-32B",
            "tooltip": "Qwen-2.5-32B-Instruct"
        },
        "url": "https://arxiv.org/abs/2502.01600",
        "date": "2025-04-09",
        "test_normal": {
            "all": {
                "task_goal_completion": 72.6,
                "scenario_goal_completion": 53.6,
                "interactions": 17.44
            },
            "level 1": {
                "task_goal_completion": 87.7,
                "scenario_goal_completion": 79.0,
                "interactions": 17.44
            },
            "level 2": {
                "task_goal_completion": 89.6,
                "scenario_goal_completion": 75.0,
                "interactions": 17.44
            },
            "level 3": {
                "task_goal_completion": 46.0,
                "scenario_goal_completion": 14.3,
                "interactions": 17.44
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 47.2,
                "scenario_goal_completion": 28.8,
                "interactions": 22.23
            },
            "level 1": {
                "task_goal_completion": 73.6,
                "scenario_goal_completion": 58.3,
                "interactions": 22.23
            },
            "level 2": {
                "task_goal_completion": 40.7,
                "scenario_goal_completion": 22.0,
                "interactions": 22.23
            },
            "level 3": {
                "task_goal_completion": 42.6,
                "scenario_goal_completion": 23.1,
                "interactions": 22.23
            }
        }
    },
    {
        "id": "8a02d49a-0df800d6",
        "method": {
            "name": "IBM CUGA",
            "tooltip": "IBM Computer Using Generalist Agent"
        },
        "llm": {
            "name": "GPT-4.1",
            "tooltip": "gpt-4.1-2025-04-14"
        },
        "url": "https://cuga.dev",
        "date": "2025-07-12",
        "test_normal": {
            "all": {
                "task_goal_completion": 73.2,
                "scenario_goal_completion": 62.5,
                "interactions": 10.69
            },
            "level 1": {
                "task_goal_completion": 91.2,
                "scenario_goal_completion": 84.2,
                "interactions": 5.94
            },
            "level 2": {
                "task_goal_completion": 77.1,
                "scenario_goal_completion": 68.8,
                "interactions": 10.36
            },
            "level 3": {
                "task_goal_completion": 54.0,
                "scenario_goal_completion": 38.1,
                "interactions": 12.69
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 57.6,
                "scenario_goal_completion": 48.2,
                "interactions": 8.4
            },
            "level 1": {
                "task_goal_completion": 91.7,
                "scenario_goal_completion": 87.5,
                "interactions": 4.65
            },
            "level 2": {
                "task_goal_completion": 58.7,
                "scenario_goal_completion": 42.0,
                "interactions": 8.33
            },
            "level 3": {
                "task_goal_completion": 44.1,
                "scenario_goal_completion": 38.5,
                "interactions": 11.86
            }
        }
    },
    {
        "id": "1c12639a-57a0ac34",
        "method": {
            "name": "ReAct + 2 SetBSR Demos",
            "tooltip": "ReAct + 2 SetBSR-Selected Trajectory Demonstrations"
        },
        "llm": {
            "name": "GPT-4o",
            "tooltip": "gpt-4o-2024-08-06"
        },
        "url": "https://arxiv.org/abs/2506.13109",
        "date": "2025-07-13",
        "test_normal": {
            "all": {
                "task_goal_completion": 68.5,
                "scenario_goal_completion": 57.1,
                "interactions": 23.85
            },
            "level 1": {
                "task_goal_completion": 87.7,
                "scenario_goal_completion": 73.7,
                "interactions": 23.85
            },
            "level 2": {
                "task_goal_completion": 79.2,
                "scenario_goal_completion": 68.8,
                "interactions": 23.85
            },
            "level 3": {
                "task_goal_completion": 42.9,
                "scenario_goal_completion": 33.3,
                "interactions": 23.85
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 38.9,
                "scenario_goal_completion": 23.0,
                "interactions": 28.59
            },
            "level 1": {
                "task_goal_completion": 68.1,
                "scenario_goal_completion": 54.2,
                "interactions": 28.59
            },
            "level 2": {
                "task_goal_completion": 34.0,
                "scenario_goal_completion": 14.0,
                "interactions": 28.59
            },
            "level 3": {
                "task_goal_completion": 31.8,
                "scenario_goal_completion": 18.5,
                "interactions": 28.59
            }
        }
    },
    {
        "id": "cabab065-0caee3c3",
        "method": {
            "name": "Alibaba Cloud ApsaraLab AgentRL",
            "tooltip": "Alibaba Cloud ApsaraLab AgentRL"
        },
        "llm": {
            "name": "Qwen3-14B",
            "tooltip": "Qwen3-14B"
        },
        "url": "https://www.aliyun.com/product/bailian",
        "date": "2026-02-15",
        "test_normal": {
            "all": {
                "task_goal_completion": 86.9,
                "scenario_goal_completion": 80.4,
                "interactions": 29.51
            },
            "level 1": {
                "task_goal_completion": 100.0,
                "scenario_goal_completion": 100.0,
                "interactions": 29.51
            },
            "level 2": {
                "task_goal_completion": 91.7,
                "scenario_goal_completion": 87.5,
                "interactions": 29.51
            },
            "level 3": {
                "task_goal_completion": 71.4,
                "scenario_goal_completion": 57.1,
                "interactions": 29.51
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 67.6,
                "scenario_goal_completion": 50.4,
                "interactions": 41.78
            },
            "level 1": {
                "task_goal_completion": 88.9,
                "scenario_goal_completion": 79.2,
                "interactions": 41.78
            },
            "level 2": {
                "task_goal_completion": 62.7,
                "scenario_goal_completion": 40.0,
                "interactions": 41.78
            },
            "level 3": {
                "task_goal_completion": 63.6,
                "scenario_goal_completion": 47.7,
                "interactions": 41.78
            }
        }
    },
    {
        "id": "5acab4bb-badb978f",
        "method": {
            "name": "RooAGI_agentbuilder_memtune_rl",
            "tooltip": "RooAGI AgentBuilder MEMTune RL"
        },
        "llm": {
            "name": "gptoss-20b",
            "tooltip": "openai"
        },
        "url": "https://agentbuilder.rooagi.com/",
        "date": "2026-02-22",
        "test_normal": {
            "all": {
                "task_goal_completion": 76.2,
                "scenario_goal_completion": 58.9,
                "interactions": 10.85
            },
            "level 1": {
                "task_goal_completion": 91.2,
                "scenario_goal_completion": 73.7,
                "interactions": 10.85
            },
            "level 2": {
                "task_goal_completion": 91.7,
                "scenario_goal_completion": 75.0,
                "interactions": 10.85
            },
            "level 3": {
                "task_goal_completion": 50.8,
                "scenario_goal_completion": 33.3,
                "interactions": 10.85
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 15.3,
                "scenario_goal_completion": 7.9,
                "interactions": 17.03
            },
            "level 1": {
                "task_goal_completion": 55.6,
                "scenario_goal_completion": 37.5,
                "interactions": 17.03
            },
            "level 2": {
                "task_goal_completion": 11.3,
                "scenario_goal_completion": 2.0,
                "interactions": 17.03
            },
            "level 3": {
                "task_goal_completion": 3.6,
                "scenario_goal_completion": 1.5,
                "interactions": 17.03
            }
        }
    },
    {
        "id": "f072cd07-6ae0fbdf",
        "method": {
            "name": "Button Agent",
            "tooltip": "Button Agent"
        },
        "llm": {
            "name": "Claude Opus 4.6",
            "tooltip": "claude-opus-4-6"
        },
        "url": "https://distyl.ai/",
        "date": "2026-03-23",
        "test_normal": {
            "all": {
                "task_goal_completion": 91.1,
                "scenario_goal_completion": 89.3,
                "interactions": 15.89
            },
            "level 1": {
                "task_goal_completion": 82.5,
                "scenario_goal_completion": 79.0,
                "interactions": 15.89
            },
            "level 2": {
                "task_goal_completion": 100.0,
                "scenario_goal_completion": 100.0,
                "interactions": 15.89
            },
            "level 3": {
                "task_goal_completion": 92.1,
                "scenario_goal_completion": 90.5,
                "interactions": 15.89
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 72.4,
                "scenario_goal_completion": 64.8,
                "interactions": 20.52
            },
            "level 1": {
                "task_goal_completion": 76.4,
                "scenario_goal_completion": 70.8,
                "interactions": 20.52
            },
            "level 2": {
                "task_goal_completion": 70.0,
                "scenario_goal_completion": 62.0,
                "interactions": 20.52
            },
            "level 3": {
                "task_goal_completion": 72.8,
                "scenario_goal_completion": 64.6,
                "interactions": 20.52
            }
        }
    },
    {
        "id": "1095b768-6513e0c6",
        "method": {
            "name": "kecaipan capybara",
            "tooltip": "Self-organizing context assets learned on train and frozen for closed-book evaluation"
        },
        "llm": {
            "name": "gpt-5.6-luna",
            "tooltip": "LongjiTech OpenAI-compatible Responses API; model ID gpt-5.6-luna"
        },
        "url": "https://github.com/pkc000pkc/capybara",
        "date": "2026-08-08",
        "test_normal": {
            "all": {
                "task_goal_completion": 85.1,
                "scenario_goal_completion": 73.2,
                "interactions": 9.34
            },
            "level 1": {
                "task_goal_completion": 94.7,
                "scenario_goal_completion": 84.2,
                "interactions": 9.34
            },
            "level 2": {
                "task_goal_completion": 93.8,
                "scenario_goal_completion": 81.2,
                "interactions": 9.34
            },
            "level 3": {
                "task_goal_completion": 69.8,
                "scenario_goal_completion": 57.1,
                "interactions": 9.34
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 73.4,
                "scenario_goal_completion": 52.5,
                "interactions": 11.17
            },
            "level 1": {
                "task_goal_completion": 83.3,
                "scenario_goal_completion": 58.3,
                "interactions": 11.17
            },
            "level 2": {
                "task_goal_completion": 68.7,
                "scenario_goal_completion": 48.0,
                "interactions": 11.17
            },
            "level 3": {
                "task_goal_completion": 73.3,
                "scenario_goal_completion": 53.8,
                "interactions": 11.17
            }
        }
    },
    {
        "id": "695a4b55-1b68dde4",
        "method": {
            "name": "LARA MAS",
            "tooltip": "Explorer writes an app-tagged plan; per-app specialist executors run a ReAct loop; single attempt per task"
        },
        "llm": {
            "name": "gpt-4.1-mini",
            "tooltip": "gpt-4.1-mini for all agents (explorer, executor, supervisor)"
        },
        "url": "https://github.com/Omer212111/LARA",
        "date": "2026-08-19",
        "test_normal": {
            "all": {
                "task_goal_completion": 61.9,
                "scenario_goal_completion": 50.0,
                "interactions": 21.11
            },
            "level 1": {
                "task_goal_completion": 91.2,
                "scenario_goal_completion": 79.0,
                "interactions": 21.11
            },
            "level 2": {
                "task_goal_completion": 52.1,
                "scenario_goal_completion": 43.8,
                "interactions": 21.11
            },
            "level 3": {
                "task_goal_completion": 42.9,
                "scenario_goal_completion": 28.6,
                "interactions": 21.11
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 37.6,
                "scenario_goal_completion": 20.1,
                "interactions": 24.75
            },
            "level 1": {
                "task_goal_completion": 72.2,
                "scenario_goal_completion": 45.8,
                "interactions": 24.75
            },
            "level 2": {
                "task_goal_completion": 36.0,
                "scenario_goal_completion": 18.0,
                "interactions": 24.75
            },
            "level 3": {
                "task_goal_completion": 26.2,
                "scenario_goal_completion": 12.3,
                "interactions": 24.75
            }
        }
    },
    {
        "id": "23715209-d5c9ade3",
        "method": {
            "name": "LARA MAS",
            "tooltip": "Explorer writes an app-tagged plan; per-app specialist executors run a ReAct loop; single attempt per task"
        },
        "llm": {
            "name": "claude-opus-4-7",
            "tooltip": "claude-opus-4-7 for all agents (explorer, executor, supervisor)"
        },
        "url": "https://github.com/Omer212111/LARA",
        "date": "2026-09-02",
        "test_normal": {
            "all": {
                "task_goal_completion": 88.7,
                "scenario_goal_completion": 82.1,
                "interactions": 18.02
            },
            "level 1": {
                "task_goal_completion": 98.2,
                "scenario_goal_completion": 94.7,
                "interactions": 18.02
            },
            "level 2": {
                "task_goal_completion": 89.6,
                "scenario_goal_completion": 81.2,
                "interactions": 18.02
            },
            "level 3": {
                "task_goal_completion": 79.4,
                "scenario_goal_completion": 71.4,
                "interactions": 18.02
            }
        },
        "test_challenge": {
            "all": {
                "task_goal_completion": 85.6,
                "scenario_goal_completion": 77.0,
                "interactions": 22.15
            },
            "level 1": {
                "task_goal_completion": 91.7,
                "scenario_goal_completion": 83.3,
                "interactions": 22.15
            },
            "level 2": {
                "task_goal_completion": 84.7,
                "scenario_goal_completion": 76.0,
                "interactions": 22.15
            },
            "level 3": {
                "task_goal_completion": 84.1,
                "scenario_goal_completion": 75.4,
                "interactions": 22.15
            }
        }
    }
]