{
  "title": "Two Models. Two Very Different Gardens.",
  "recording": {
    "source_filename": "pagoda_duel_send.mp4",
    "source_sha256": "bebeb11695cc5552314c35c119596bea7c4199a0e459819d6c6aa5def2ac5dd1",
    "output_sha256": "86c9b5cec24a47efdbc52c1308e29e2fe95ce8d1771484ce94190dbddb60bbf9",
    "edit": "Scene-only crop; historical stat board removed. Spatial downscale and H264 re-encode; source cadence retained.",
    "crop": "1920:720:0:100",
    "probe": {
      "programs": [],
      "stream_groups": [],
      "streams": [
        {
          "width": 1280,
          "height": 480,
          "avg_frame_rate": "30/1"
        }
      ],
      "format": {
        "duration": "60.000000",
        "size": "13779680"
      }
    }
  },
  "arms": [
    {
      "label": "Qwen Flash-Next FP8 \u00b7 vLLM TP2",
      "wall_s": 2073.2,
      "completion_tokens": 82830,
      "reasoning_chars": 143364,
      "content_chars": 81801,
      "finish": "stop",
      "settings": {
        "thinking": true,
        "reasoning_effort": "xhigh",
        "temperature": 0.7,
        "max_tokens": 120000,
        "stream": true
      }
    },
    {
      "label": "Ling-3.0 Flash VL FP8 \u00b7 SGLang TP2",
      "wall_s": 862.2,
      "completion_tokens": 22139,
      "reasoning_chars": 37221,
      "content_chars": 24679,
      "finish": "stop",
      "settings": {
        "thinking": true,
        "reasoning_effort": "xhigh",
        "temperature": 0.7,
        "max_tokens": 120000,
        "stream": true
      }
    }
  ],
  "artifact_hashes": [
    {
      "file": "tp2-qwen-next.html",
      "sha256": "dfd1a794883fba882653645a0aac8b08916cc68d560c67ec022062cc75cf076a",
      "bytes": 83848
    },
    {
      "file": "ling-vl-fp8.html",
      "sha256": "677871fe4e1f14b2841c3117445baf553a11858610c8c1792822fecc14f654ab",
      "bytes": 24675
    }
  ],
  "notes": [
    "Both requests enabled thinking, requested xhigh, used temperature 0.7 and an explicit 120,000-token output budget. An identical effort label does not establish equivalent reasoning policy across model families.",
    "Both used two DGX Sparks, but different model architectures and serving engines. Qwen\u2019s recipe used MTP; Ling\u2019s did not. No HTML repairs are recorded for the displayed pagoda artifacts.",
    "Reasoning was recorded as characters in the run metadata, not a verified per-arm reasoning-token split. Completion-token counts include reasoning. The video is a cropped excerpt of the existing comparison.",
    "The Ling output is the same saved generation used in both the Qwen vs Ling and Nex vs Ling comparisons, not an independent second sample."
  ]
}