{
  "apiVersion": "1",
  "methodologyVersion": "2026-07-09-gso-ml-domain",
  "updatedAt": "2026-07-13T21:57:32.936Z",
  "id": "cybersecurity",
  "name": "Cybersecurity",
  "url": "https://www.aiiq.org/cybersecurity/",
  "models": [
    {
      "model": "gpt-5.5",
      "compositeIQ": 147,
      "benchmarksCovered": 4,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
    },
    {
      "model": "mythos-5",
      "compositeIQ": 147,
      "benchmarksCovered": 3,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "glm-5.2",
      "compositeIQ": 139,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/glm-5.2/"
    },
    {
      "model": "gpt-5.4",
      "compositeIQ": 138,
      "benchmarksCovered": 4,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gpt-5.4/"
    },
    {
      "model": "minimax-m3",
      "compositeIQ": 138,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/minimax-m3/"
    },
    {
      "model": "gpt-5.5-cyber",
      "compositeIQ": 137,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "deepseek-v4-pro",
      "compositeIQ": 131,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/deepseek-v4-pro/"
    },
    {
      "model": "gemini-3.5-flash",
      "compositeIQ": 131,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gemini-3.5-flash/"
    },
    {
      "model": "multi-model-mdash",
      "compositeIQ": 131,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "opus-4.6",
      "compositeIQ": 131,
      "benchmarksCovered": 3,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/opus-4.6/"
    },
    {
      "model": "deepseek-v4-flash",
      "compositeIQ": 130,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/deepseek-v4-flash/"
    },
    {
      "model": "gemini-3.1-pro",
      "compositeIQ": 121,
      "benchmarksCovered": 4,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gemini-3.1-pro/"
    },
    {
      "model": "opus-4.8",
      "compositeIQ": 121,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/opus-4.8/"
    },
    {
      "model": "grok-4.3",
      "compositeIQ": 117,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/grok-4.3/"
    },
    {
      "model": "kimi-k2.6",
      "compositeIQ": 117,
      "benchmarksCovered": 2,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/kimi-k2.6/"
    },
    {
      "model": "opus-4.7",
      "compositeIQ": 117,
      "benchmarksCovered": 3,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/opus-4.7/"
    },
    {
      "model": "sonnet-4.6",
      "compositeIQ": 117,
      "benchmarksCovered": 2,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/sonnet-4.6/"
    },
    {
      "model": "opus-4.5",
      "compositeIQ": 114,
      "benchmarksCovered": 2,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/opus-4.5/"
    },
    {
      "model": "gpt-5",
      "compositeIQ": 113,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gpt-5/"
    },
    {
      "model": "kimi-k2.5",
      "compositeIQ": 113,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/kimi-k2.5/"
    },
    {
      "model": "muse-spark",
      "compositeIQ": 113,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/muse-spark/"
    },
    {
      "model": "glm-5",
      "compositeIQ": 112,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/glm-5/"
    },
    {
      "model": "qwen3.7-plus",
      "compositeIQ": 112,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/qwen3.7-plus/"
    },
    {
      "model": "gemini-3-pro",
      "compositeIQ": 111,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gemini-3-pro/"
    },
    {
      "model": "glm-5.1",
      "compositeIQ": 111,
      "benchmarksCovered": 3,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/glm-5.1/"
    },
    {
      "model": "gpt-5.2",
      "compositeIQ": 110,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/gpt-5.2/"
    },
    {
      "model": "opus-4.1",
      "compositeIQ": 106,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/opus-4.1/"
    },
    {
      "model": "sonnet-4.5",
      "compositeIQ": 106,
      "benchmarksCovered": 2,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "sonnet-4",
      "compositeIQ": 105,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "sonnet-3.7",
      "compositeIQ": 102,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "haiku-4.5",
      "compositeIQ": 99,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/haiku-4.5/"
    },
    {
      "model": "minimax-m2.7",
      "compositeIQ": 99,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": "https://www.aiiq.org/models/minimax-m2.7/"
    },
    {
      "model": "gpt-4.1",
      "compositeIQ": 97,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "gemini-2.5-flash",
      "compositeIQ": 85,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    },
    {
      "model": "deepseek-v3",
      "compositeIQ": 81,
      "benchmarksCovered": 1,
      "benchmarksTotal": 5,
      "modelUrl": null
    }
  ],
  "benchmarks": [
    {
      "id": "cyberbench",
      "name": "CyberBench",
      "description": "Overall success across VALS CyberBench PoC and Patch tracks for OSS-Fuzz crash reproduction and repair.",
      "direction": "higher_is_better",
      "unit": "percent",
      "source": {
        "name": "VALS CyberBench",
        "url": "https://www.vals.ai/benchmarks/cyber"
      },
      "results": [
        {
          "model": "gpt-5.5",
          "harness": "mini-swe-agent",
          "score": 80.508,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "glm-5.2",
          "harness": "mini-swe-agent",
          "score": 77.359,
          "modelUrl": "https://www.aiiq.org/models/glm-5.2/"
        },
        {
          "model": "minimax-m3",
          "harness": "mini-swe-agent",
          "score": 76.271,
          "modelUrl": "https://www.aiiq.org/models/minimax-m3/"
        },
        {
          "model": "kimi-k2.6",
          "harness": "mini-swe-agent",
          "score": 74.351,
          "modelUrl": "https://www.aiiq.org/models/kimi-k2.6/"
        },
        {
          "model": "gpt-5.4",
          "harness": "mini-swe-agent",
          "score": 72.881,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.4/"
        },
        {
          "model": "gemini-3.5-flash",
          "harness": "mini-swe-agent",
          "score": 69.331,
          "modelUrl": "https://www.aiiq.org/models/gemini-3.5-flash/"
        },
        {
          "model": "deepseek-v4-pro",
          "harness": "mini-swe-agent",
          "score": 68.644,
          "modelUrl": "https://www.aiiq.org/models/deepseek-v4-pro/"
        },
        {
          "model": "deepseek-v4-flash",
          "harness": "mini-swe-agent",
          "score": 66.716,
          "modelUrl": "https://www.aiiq.org/models/deepseek-v4-flash/"
        },
        {
          "model": "opus-4.7",
          "harness": "mini-swe-agent",
          "score": 52.141,
          "modelUrl": "https://www.aiiq.org/models/opus-4.7/"
        },
        {
          "model": "opus-4.8",
          "harness": "mini-swe-agent",
          "score": 50.847,
          "modelUrl": "https://www.aiiq.org/models/opus-4.8/"
        },
        {
          "model": "grok-4.3",
          "harness": "mini-swe-agent",
          "score": 47.457,
          "modelUrl": "https://www.aiiq.org/models/grok-4.3/"
        },
        {
          "model": "qwen3.7-plus",
          "harness": "mini-swe-agent",
          "score": 37.5,
          "modelUrl": "https://www.aiiq.org/models/qwen3.7-plus/"
        },
        {
          "model": "gemini-3.1-pro",
          "harness": "mini-swe-agent",
          "score": 36.441,
          "modelUrl": "https://www.aiiq.org/models/gemini-3.1-pro/"
        }
      ]
    },
    {
      "id": "exploitbench",
      "name": "ExploitBench",
      "description": "Cap coverage on the ExploitBench capability ladder for V8 exploitation.",
      "direction": "higher_is_better",
      "unit": "percent",
      "source": {
        "name": "ExploitBench leaderboard",
        "url": "https://exploitbench.ai/"
      },
      "results": [
        {
          "model": "mythos-5",
          "harness": "AutoNudge",
          "score": 78,
          "modelUrl": null
        },
        {
          "model": "gpt-5.5",
          "harness": "Codex AutoNudge",
          "score": 72,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "mythos-5",
          "harness": "Baseline",
          "score": 72,
          "modelUrl": null
        },
        {
          "model": "gpt-5.5",
          "harness": "AutoNudge",
          "score": 47,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "gpt-5.5",
          "harness": "AutoNudge",
          "score": 42,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "gpt-5.5",
          "harness": "Codex",
          "score": 33,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "opus-4.7",
          "harness": "Baseline",
          "score": 28,
          "modelUrl": "https://www.aiiq.org/models/opus-4.7/"
        },
        {
          "model": "gemini-3.1-pro",
          "harness": "Baseline",
          "score": 26,
          "modelUrl": "https://www.aiiq.org/models/gemini-3.1-pro/"
        },
        {
          "model": "sonnet-4.6",
          "harness": "Baseline",
          "score": 26,
          "modelUrl": "https://www.aiiq.org/models/sonnet-4.6/"
        },
        {
          "model": "opus-4.7",
          "harness": "Baseline",
          "score": 25,
          "modelUrl": "https://www.aiiq.org/models/opus-4.7/"
        },
        {
          "model": "sonnet-4.6",
          "harness": "Baseline",
          "score": 25,
          "modelUrl": "https://www.aiiq.org/models/sonnet-4.6/"
        },
        {
          "model": "kimi-k2.6",
          "harness": "AutoNudge",
          "score": 18,
          "modelUrl": "https://www.aiiq.org/models/kimi-k2.6/"
        },
        {
          "model": "glm-5.1",
          "harness": "AutoNudge",
          "score": 18,
          "modelUrl": "https://www.aiiq.org/models/glm-5.1/"
        },
        {
          "model": "kimi-k2.6",
          "harness": "Baseline",
          "score": 16,
          "modelUrl": "https://www.aiiq.org/models/kimi-k2.6/"
        },
        {
          "model": "glm-5.1",
          "harness": "Baseline",
          "score": 16,
          "modelUrl": "https://www.aiiq.org/models/glm-5.1/"
        },
        {
          "model": "gemini-3.1-pro",
          "harness": "AutoNudge",
          "score": 16,
          "modelUrl": "https://www.aiiq.org/models/gemini-3.1-pro/"
        },
        {
          "model": "haiku-4.5",
          "harness": "AutoNudge",
          "score": 14,
          "modelUrl": "https://www.aiiq.org/models/haiku-4.5/"
        },
        {
          "model": "haiku-4.5",
          "harness": "Baseline",
          "score": 14,
          "modelUrl": "https://www.aiiq.org/models/haiku-4.5/"
        },
        {
          "model": "minimax-m2.7",
          "harness": "Baseline",
          "score": 13,
          "modelUrl": "https://www.aiiq.org/models/minimax-m2.7/"
        },
        {
          "model": "minimax-m2.7",
          "harness": "AutoNudge",
          "score": 13,
          "modelUrl": "https://www.aiiq.org/models/minimax-m2.7/"
        }
      ]
    },
    {
      "id": "cybergym",
      "name": "CyberGym",
      "description": "Success rate on CyberGym Level 1 vulnerability reproduction tasks.",
      "direction": "higher_is_better",
      "unit": "percent",
      "source": {
        "name": "CyberGym leaderboard",
        "url": "https://www.cybergym.io/cybergym/"
      },
      "results": [
        {
          "model": "opus-4.6",
          "harness": "Crystalline",
          "score": 89.6,
          "modelUrl": "https://www.aiiq.org/models/opus-4.6/"
        },
        {
          "model": "multi-model-mdash",
          "harness": "MDASH",
          "score": 88.45,
          "modelUrl": null
        },
        {
          "model": "gpt-5.5-cyber",
          "harness": "OpenAI Agent",
          "score": 85.6,
          "modelUrl": null
        },
        {
          "model": "mythos-5",
          "harness": "Anthropic Agent",
          "score": 83.1,
          "modelUrl": null
        },
        {
          "model": "gpt-5.5",
          "harness": "OpenAI Agent",
          "score": 81.8,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "gpt-5.4",
          "harness": "OpenAI Agent",
          "score": 79,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.4/"
        },
        {
          "model": "glm-5.1",
          "harness": "Claude Code",
          "score": 68.7,
          "modelUrl": "https://www.aiiq.org/models/glm-5.1/"
        },
        {
          "model": "opus-4.6",
          "harness": "Anthropic Agent",
          "score": 66.6,
          "modelUrl": "https://www.aiiq.org/models/opus-4.6/"
        },
        {
          "model": "sonnet-4.6",
          "harness": "Anthropic Agent",
          "score": 65.2,
          "modelUrl": "https://www.aiiq.org/models/sonnet-4.6/"
        },
        {
          "model": "opus-4.5",
          "harness": "Anthropic Agent",
          "score": 50.63,
          "modelUrl": "https://www.aiiq.org/models/opus-4.5/"
        },
        {
          "model": "muse-spark",
          "harness": "Meta Agent",
          "score": 43.5,
          "modelUrl": "https://www.aiiq.org/models/muse-spark/"
        },
        {
          "model": "glm-5",
          "harness": "Claude Code",
          "score": 43.2,
          "modelUrl": "https://www.aiiq.org/models/glm-5/"
        },
        {
          "model": "kimi-k2.5",
          "harness": "Kimi Agent",
          "score": 41.3,
          "modelUrl": "https://www.aiiq.org/models/kimi-k2.5/"
        },
        {
          "model": "gpt-5",
          "harness": "OpenHands",
          "score": 39.4,
          "modelUrl": "https://www.aiiq.org/models/gpt-5/"
        },
        {
          "model": "sonnet-4.5",
          "harness": "Anthropic Agent",
          "score": 28.9,
          "modelUrl": null
        },
        {
          "model": "opus-4.1",
          "harness": "Anthropic Agent",
          "score": 25,
          "modelUrl": "https://www.aiiq.org/models/opus-4.1/"
        },
        {
          "model": "sonnet-4",
          "harness": "Anthropic Agent",
          "score": 22.6,
          "modelUrl": null
        },
        {
          "model": "sonnet-3.7",
          "harness": "Anthropic Agent",
          "score": 14.5,
          "modelUrl": null
        },
        {
          "model": "gpt-4.1",
          "harness": "OpenHands",
          "score": 9.356,
          "modelUrl": null
        },
        {
          "model": "gemini-2.5-flash",
          "harness": "OpenHands",
          "score": 4.844,
          "modelUrl": null
        },
        {
          "model": "deepseek-v3",
          "harness": "OpenHands",
          "score": 3.583,
          "modelUrl": null
        }
      ]
    },
    {
      "id": "exploitgym",
      "name": "ExploitGym",
      "description": "Intended-vulnerability exploits completed across userspace, V8, and Linux kernel tasks.",
      "direction": "higher_is_better",
      "unit": "score",
      "source": {
        "name": "ExploitGym leaderboard",
        "url": "https://www.cybergym.io/exploitgym/"
      },
      "results": [
        {
          "model": "mythos-5",
          "harness": "Claude Code",
          "score": 157,
          "modelUrl": null
        },
        {
          "model": "gpt-5.5",
          "harness": "Codex CLI",
          "score": 131,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.5/"
        },
        {
          "model": "gpt-5.4",
          "harness": "Codex CLI",
          "score": 61,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.4/"
        },
        {
          "model": "opus-4.6",
          "harness": "Claude Code",
          "score": 16,
          "modelUrl": "https://www.aiiq.org/models/opus-4.6/"
        },
        {
          "model": "gemini-3.1-pro",
          "harness": "Gemini CLI",
          "score": 12,
          "modelUrl": "https://www.aiiq.org/models/gemini-3.1-pro/"
        },
        {
          "model": "opus-4.7",
          "harness": "Claude Code",
          "score": 12,
          "modelUrl": "https://www.aiiq.org/models/opus-4.7/"
        },
        {
          "model": "glm-5.1",
          "harness": "Claude Code",
          "score": 4,
          "modelUrl": "https://www.aiiq.org/models/glm-5.1/"
        }
      ]
    },
    {
      "id": "cybergym-e2e",
      "name": "CyberGym-E2E",
      "description": "S3 end-to-end discover-and-patch success rate: PoC crashes, patch fixes it, and tests pass.",
      "direction": "higher_is_better",
      "unit": "percent",
      "source": {
        "name": "CyberGym-E2E leaderboard",
        "url": "https://www.cybergym.io/cybergym-e2e/"
      },
      "results": [
        {
          "model": "gpt-5.4",
          "harness": "Codex",
          "score": 65.9,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.4/"
        },
        {
          "model": "opus-4.6",
          "harness": "Claude Code",
          "score": 62.6,
          "modelUrl": "https://www.aiiq.org/models/opus-4.6/"
        },
        {
          "model": "gemini-3.1-pro",
          "harness": "Gemini CLI",
          "score": 43.8,
          "modelUrl": "https://www.aiiq.org/models/gemini-3.1-pro/"
        },
        {
          "model": "opus-4.6",
          "harness": "Claude Code",
          "score": 37.9,
          "modelUrl": "https://www.aiiq.org/models/opus-4.6/"
        },
        {
          "model": "gemini-3-pro",
          "harness": "Gemini CLI",
          "score": 22.6,
          "modelUrl": "https://www.aiiq.org/models/gemini-3-pro/"
        },
        {
          "model": "gpt-5.2",
          "harness": "Codex",
          "score": 20.7,
          "modelUrl": "https://www.aiiq.org/models/gpt-5.2/"
        },
        {
          "model": "opus-4.5",
          "harness": "Claude Code",
          "score": 19.2,
          "modelUrl": "https://www.aiiq.org/models/opus-4.5/"
        },
        {
          "model": "sonnet-4.5",
          "harness": "Claude Code",
          "score": 10.6,
          "modelUrl": null
        },
        {
          "model": "sonnet-4.5",
          "harness": "OpenHands",
          "score": 5.4,
          "modelUrl": null
        }
      ]
    }
  ]
}