{
  "url": "https://www.aibattle.space/rankings/eval-hle-humanity-s-last-exam-cost-per-task",
  "dataset": "Humanity's Last Exam: Cost per Task",
  "family": "tab::eval-hle",
  "familyLabel": "Frontier Evals",
  "sourceRowCount": 20,
  "window": 20,
  "metricKeys": [
    "answer",
    "cacheHit",
    "cacheWrite",
    "input",
    "reasoning"
  ],
  "rebuilt": "2026-09-19",
  "stats": {
    "median": 0.00028121802594995363,
    "mean": 0.0006420583716347314,
    "min": 0.00004665556070435588,
    "max": 0.004317295611563948,
    "spread": 0.004270640050859593,
    "gapFirstToSecond": 0.00003656156163113995,
    "gapFirstToSecondPercent": 43.94,
    "creators": 12
  },
  "entries": [
    {
      "rank": 1,
      "name": "gpt-oss-120b (high)",
      "creator": "OpenAI",
      "slug": "gpt-oss-120b",
      "detailsUrl": "/models/gpt-oss-120b",
      "value": 0.00004665556070435588
    },
    {
      "rank": 2,
      "name": "Muse Glimmer (high)",
      "creator": "Meta",
      "slug": "muse-glimmer",
      "detailsUrl": "/models/muse-glimmer",
      "value": 0.00008321712233549583
    },
    {
      "rank": 3,
      "name": "GLM-5.3-Flash",
      "creator": "Z AI",
      "slug": "glm-5-3-flash",
      "detailsUrl": "/models/glm-5-3-flash",
      "value": 0.00004794633920296571
    },
    {
      "rank": 4,
      "name": "MiniMax-M3",
      "creator": "MiniMax",
      "slug": "minimax-m3",
      "detailsUrl": "/models/minimax-m3",
      "value": 0.00013882771084337348
    },
    {
      "rank": 5,
      "name": "Gemini 3.5 Flash-Lite",
      "creator": "Google",
      "slug": "gemini-3-5-flash-lite",
      "detailsUrl": "/models/gemini-3-5-flash-lite",
      "value": 0.00009598220574606117
    },
    {
      "rank": 6,
      "name": "GPT-5.6 Luna (max)",
      "creator": "OpenAI",
      "slug": "gpt-5-6-luna",
      "detailsUrl": "/models/gpt-5-6-luna",
      "value": 0.00006208915662650603
    },
    {
      "rank": 7,
      "name": "DeepSeek V4.1 Flash (Reasoning, Max Effort)",
      "creator": "DeepSeek",
      "slug": "deepseek-v4-1-flash",
      "detailsUrl": "/models/deepseek-v4-1-flash",
      "value": 0.00009827154772937905
    },
    {
      "rank": 8,
      "name": "Nemotron 3 Ultra 550B A55B (Reasoning)",
      "creator": "NVIDIA",
      "slug": "nvidia-nemotron-3-ultra-550b-a55b",
      "detailsUrl": "/models/nvidia-nemotron-3-ultra-550b-a55b",
      "value": 0.00019808758109360515
    },
    {
      "rank": 9,
      "name": "Gemini 3.8 Flash (high)",
      "creator": "Google",
      "slug": "gemini-3-8-flash",
      "detailsUrl": "/models/gemini-3-8-flash",
      "value": 0.0002399277108433735
    },
    {
      "rank": 10,
      "name": "Grok 4.6 (high)",
      "creator": "SpaceXAI",
      "slug": "grok-4-6",
      "detailsUrl": "/models/grok-4-6",
      "value": 0.0006642409638554217
    },
    {
      "rank": 11,
      "name": "Muse Spark 1.3 (max)",
      "creator": "Meta",
      "slug": "muse-spark-1-3",
      "detailsUrl": "/models/muse-spark-1-3",
      "value": 0.00038077386468952733
    },
    {
      "rank": 12,
      "name": "Qwen3.8 27B (xhigh)",
      "creator": "Alibaba",
      "slug": "qwen3-8-27b",
      "detailsUrl": "/models/qwen3-8-27b",
      "value": 0.00019132414272474515
    },
    {
      "rank": 13,
      "name": "DeepSeek V4 Pro 0813 (Reasoning, Max Effort)",
      "creator": "DeepSeek",
      "slug": "deepseek-v4-pro",
      "detailsUrl": "/models/deepseek-v4-pro",
      "value": 0.0005194591473586654
    },
    {
      "rank": 14,
      "name": "Inkling (xhigh)",
      "creator": "Thinking Machines",
      "slug": "inkling",
      "detailsUrl": "/models/inkling",
      "value": 0.0003225083410565338
    },
    {
      "rank": 15,
      "name": "Qwen3.8 2.4T A95B",
      "creator": "Alibaba",
      "slug": "qwen3-8-2-4t-a95b",
      "detailsUrl": "/models/qwen3-8-2-4t-a95b",
      "value": 0.0007523253012048193
    },
    {
      "rank": 16,
      "name": "Mistral Medium 3.5",
      "creator": "Mistral",
      "slug": "mistral-medium-3-5",
      "detailsUrl": "/models/mistral-medium-3-5",
      "value": 0.00046383225208526414
    },
    {
      "rank": 17,
      "name": "Claude Fable 5.1 (Adaptive Reasoning, Medium Effort, Default Fallback)",
      "creator": "Anthropic",
      "slug": "claude-fable-5-1-medium",
      "detailsUrl": "/models/claude-fable-5-1-medium",
      "value": 0.004317295611563948
    },
    {
      "rank": 18,
      "name": "GPT-5.6 Terra (max)",
      "creator": "OpenAI",
      "slug": "gpt-5-6-terra",
      "detailsUrl": "/models/gpt-5-6-terra",
      "value": 0.0006130157553290084
    },
    {
      "rank": 19,
      "name": "GLM-5.3 (max)",
      "creator": "Z AI",
      "slug": "glm-5-3",
      "detailsUrl": "/models/glm-5-3",
      "value": 0.00045030370713623725
    },
    {
      "rank": 20,
      "name": "GPT-6 Astra (xhigh)",
      "creator": "OpenAI",
      "slug": "gpt-6-astra-xhigh",
      "detailsUrl": "/models/gpt-6-astra-xhigh",
      "value": 0.003155083410565338
    }
  ]
}