{
  "url": "https://www.aibattle.space/rankings/eval-gdpval-gdpval-aa-v2-output-tokens-per-task",
  "dataset": "GDPval-AA v2: Output Tokens per Task",
  "family": "tab::eval-gdpval",
  "familyLabel": "eval gdpval",
  "sourceRowCount": 20,
  "window": 20,
  "metricKeys": [
    "answer",
    "reasoning"
  ],
  "rebuilt": "2026-09-19",
  "stats": {
    "median": 24291.81590909091,
    "mean": 25024.530000000002,
    "min": 5332.631818181818,
    "max": 45717.740909090906,
    "spread": 40385.109090909085,
    "gapFirstToSecond": 11080.086363636365,
    "gapFirstToSecondPercent": 63.13,
    "creators": 12
  },
  "entries": [
    {
      "rank": 1,
      "name": "Muse Glimmer (high)",
      "creator": "Meta",
      "slug": "muse-glimmer",
      "detailsUrl": "/models/muse-glimmer",
      "value": 6471.113636363636
    },
    {
      "rank": 2,
      "name": "Mistral Medium 3.5",
      "creator": "Mistral",
      "slug": "mistral-medium-3-5",
      "detailsUrl": "/models/mistral-medium-3-5",
      "value": 17551.2
    },
    {
      "rank": 3,
      "name": "Gemini 3.5 Flash-Lite",
      "creator": "Google",
      "slug": "gemini-3-5-flash-lite",
      "detailsUrl": "/models/gemini-3-5-flash-lite",
      "value": 15236.709090909091
    },
    {
      "rank": 4,
      "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": 23544.736363636363
    },
    {
      "rank": 5,
      "name": "Inkling (xhigh)",
      "creator": "Thinking Machines",
      "slug": "inkling",
      "detailsUrl": "/models/inkling",
      "value": 19971.418181818182
    },
    {
      "rank": 6,
      "name": "gpt-oss-120b (high)",
      "creator": "OpenAI",
      "slug": "gpt-oss-120b",
      "detailsUrl": "/models/gpt-oss-120b",
      "value": 5332.631818181818
    },
    {
      "rank": 7,
      "name": "GPT-5.6 Luna (max)",
      "creator": "OpenAI",
      "slug": "gpt-5-6-luna",
      "detailsUrl": "/models/gpt-5-6-luna",
      "value": 18049.25909090909
    },
    {
      "rank": 8,
      "name": "Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback)",
      "creator": "Anthropic",
      "slug": "claude-fable-5-1-high",
      "detailsUrl": "/models/claude-fable-5-1-high",
      "value": 27972.922727272726
    },
    {
      "rank": 9,
      "name": "GPT-6 Astra (max)",
      "creator": "OpenAI",
      "slug": "gpt-6-astra",
      "detailsUrl": "/models/gpt-6-astra",
      "value": 18050.313636363637
    },
    {
      "rank": 10,
      "name": "GPT-5.6 Sol (max)",
      "creator": "OpenAI",
      "slug": "gpt-5-6-sol",
      "detailsUrl": "/models/gpt-5-6-sol",
      "value": 22677.1
    },
    {
      "rank": 11,
      "name": "GPT-5.6 Terra (max)",
      "creator": "OpenAI",
      "slug": "gpt-5-6-terra",
      "detailsUrl": "/models/gpt-5-6-terra",
      "value": 20050.445454545454
    },
    {
      "rank": 12,
      "name": "MiniMax-M3",
      "creator": "MiniMax",
      "slug": "minimax-m3",
      "detailsUrl": "/models/minimax-m3",
      "value": 31653.713636363635
    },
    {
      "rank": 13,
      "name": "Grok 4.6 (high)",
      "creator": "SpaceXAI",
      "slug": "grok-4-6",
      "detailsUrl": "/models/grok-4-6",
      "value": 36001.84090909091
    },
    {
      "rank": 14,
      "name": "Kimi K3 (max)",
      "creator": "Kimi",
      "slug": "kimi-k3",
      "detailsUrl": "/models/kimi-k3",
      "value": 25869.977272727272
    },
    {
      "rank": 15,
      "name": "Grok 4.6 (xhigh)",
      "creator": "SpaceXAI",
      "slug": "grok-4-6-xhigh",
      "detailsUrl": "/models/grok-4-6-xhigh",
      "value": 39451.75
    },
    {
      "rank": 16,
      "name": "Muse Spark 1.3 (xhigh)",
      "creator": "Meta",
      "slug": "muse-spark-1-3-xhigh",
      "detailsUrl": "/models/muse-spark-1-3-xhigh",
      "value": 45717.740909090906
    },
    {
      "rank": 17,
      "name": "K2 Horizon 375B A23B",
      "creator": "MBZUAI Institute of Foundation Models",
      "slug": "k2-horizon-375b-a23b",
      "detailsUrl": "/models/k2-horizon-375b-a23b",
      "value": 25038.895454545454
    },
    {
      "rank": 18,
      "name": "DeepSeek V4 Pro 0813 (Reasoning, Max Effort)",
      "creator": "DeepSeek",
      "slug": "deepseek-v4-pro",
      "detailsUrl": "/models/deepseek-v4-pro",
      "value": 26063.95909090909
    },
    {
      "rank": 19,
      "name": "Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)",
      "creator": "Anthropic",
      "slug": "claude-opus-5-xhigh",
      "detailsUrl": "/models/claude-opus-5-xhigh",
      "value": 44453.82272727273
    },
    {
      "rank": 20,
      "name": "Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)",
      "creator": "Anthropic",
      "slug": "claude-fable-5",
      "detailsUrl": "/models/claude-fable-5",
      "value": 31331.05
    }
  ]
}