{
  "arm": "ARM 1 / styles01",
  "suite": "Classics",
  "metrics": {
    "gsm8k / flexible-extract": {
      "score": 94.31387414708112,
      "n": 1319,
      "stderr": 0.6378790242099631,
      "unit": "%",
      "status": "verified"
    },
    "math500 / math_verify": {
      "score": 60.0,
      "n": 500,
      "stderr": 2.1930844120728583,
      "unit": "%",
      "status": "verified"
    },
    "ifeval / prompt-strict": {
      "score": 51.386321626617374,
      "n": 541,
      "stderr": 2.15083020678561,
      "unit": "%",
      "status": "verified"
    },
    "humaneval": {
      "score": null,
      "n": 0,
      "unit": "%",
      "status": "re-running",
      "note": "Pending re-run with max_gen_toks=8192; invalid 1024-token coding/ARC results excluded."
    },
    "mbpp": {
      "score": null,
      "n": 0,
      "unit": "%",
      "status": "re-running",
      "note": "Pending re-run with max_gen_toks=8192; invalid 1024-token coding/ARC results excluded."
    },
    "ARC": {
      "score": null,
      "n": 0,
      "unit": "%",
      "status": "re-running",
      "note": "Pending re-run with max_gen_toks=8192; invalid 1024-token coding/ARC results excluded."
    }
  },
  "config": {
    "thinking": true,
    "max_gen_toks": 1024,
    "temperature": 0
  },
  "evidence": {
    "engine": "styles01 vLLM + TensorFlash",
    "image": "ghcr.io/styles01/dsv41-1017-exl3-2x-dgx-sparks:v2",
    "kv": "2.68 GB / rank",
    "context": "262K server context; vision on",
    "source": "Fresh ct-bench lm-eval artifact",
    "raw": "/evidence/classics-1024.json",
    "note": "Completed 1024-token run; harness context 16,384. Displayed scores and stderr are percentages. Coding and ARC are re-running at 8192, not zero."
  },
  "ts": "2026-10-21T02:18:05.924996Z",
  "hardware": "2 \u00d7 NVIDIA DGX Spark / TP2 / EXL3 2.9bpw"
}