{
  "generated": "2026-09-08",
  "benchmark": "frontiermath",
  "results": [
    {
      "system": "gpt-4o-2024-11-20",
      "developer": "OpenAI",
      "value": 0.3,
      "date": "2025-03-06",
      "source": {
        "url": "https://epoch.ai/benchmarks/frontiermath-tiers-1-3-v1",
        "kind": "independent-evaluation",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "Tiers 1-3 (v1)",
        "tools": true,
        "notes": "Epoch AI Benchmarking Hub run of 2025-03-06 (data export); the Nov 2024 paper reported all models under 2%."
      }
    },
    {
      "system": "gpt-5.4-pro-2026-03-05 (xhigh)",
      "developer": "OpenAI",
      "value": 50.0,
      "date": "2026-03-06",
      "source": {
        "url": "https://epoch.ai/benchmarks/frontiermath-tiers-1-3-v1",
        "kind": "independent-evaluation",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "Tiers 1-3 (v1)",
        "tools": true,
        "notes": "Epoch run on v1 problem set; the highest v1 score was gpt-5.5-pro (high) at 52.4 on 2026-04-23."
      }
    },
    {
      "system": "claude-fable-5 (max)",
      "developer": "Anthropic",
      "value": 87.0,
      "date": "2026-06-09",
      "source": {
        "url": "https://epoch.ai/benchmarks/frontiermath-tiers-1-3-v2",
        "kind": "independent-evaluation",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "Tiers 1-3 (v2)",
        "tools": true,
        "notes": "First v2 run after the 2026-06-12 correction (Epoch data export, started 2026-06-09)."
      }
    },
    {
      "system": "gpt-5.6-sol (max)",
      "developer": "OpenAI",
      "value": 89.1,
      "date": "2026-07-09",
      "source": {
        "url": "https://epoch.ai/benchmarks/frontiermath-tiers-1-3-v2",
        "kind": "independent-evaluation",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "Tiers 1-3 (v2)",
        "tools": true,
        "notes": "Epoch AI Benchmarking Hub, stderr 1.8."
      }
    },
    {
      "system": "gpt-6-astra (max)",
      "developer": "OpenAI",
      "value": 93.7,
      "date": "2026-09-03",
      "source": {
        "url": "https://epoch.ai/benchmarks/frontiermath-tiers-1-3-v2",
        "kind": "independent-evaluation",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "split": "Tiers 1-3 (v2)",
        "tools": true,
        "notes": "Epoch AI Benchmarking Hub (run started 2026-08-30, published with the model on 2026-09-03), stderr 1.4; Tier 4 (v2): 95.1 at max, 97.6 at medium effort."
      }
    }
  ]
}