{
  "label": "GLM-5.3-Flash UD-Q2_K_XL (108.7 GB) reasoning HIGH",
  "tests": {
    "01-blockfall": {
      "error": "DNF: exceeded the 60 minute time limit (partial output kept: 23818 chunks, 69095 thinking chars, 13393 answer chars)",
      "partial_thinking_chars": 69095,
      "partial_answer_chars": 13393,
      "partial_chunks": 23818
    },
    "03-eruption": {
      "error": "DNF: exceeded the 60 minute time limit (partial output kept: 23805 chunks, 85787 thinking chars, 0 answer chars)",
      "partial_thinking_chars": 85787,
      "partial_answer_chars": 0,
      "partial_chunks": 23805
    },
    "04-the-ledger": {
      "error": "DNF: exceeded the 60 minute time limit (partial output kept: 23362 chunks, 55959 thinking chars, 23491 answer chars)",
      "partial_thinking_chars": 55959,
      "partial_answer_chars": 23491,
      "partial_chunks": 23362
    },
    "05-blind-artist": {
      "error": "DNF: exceeded the 60 minute time limit (partial output kept: 23703 chunks, 71042 thinking chars, 0 answer chars)",
      "partial_thinking_chars": 71042,
      "partial_answer_chars": 0,
      "partial_chunks": 23703
    }
  },
  "vision": {},
  "size_gb_on_disk": 108.7,
  "load_seconds": 30.0,
  "warmup_seconds": 3.0,
  "mem_baseline_gb": 4.52,
  "mem_peak_gb": 109.15,
  "mem_model_gb_est": 104.63,
  "backend": "llama-server direct: ~/repo/tools/llama.cpp-glm5next/build/bin/llama-server",
  "server_timings": [
    {
      "task": 0,
      "prompt_ms": 2221.66,
      "prompt_tokens": 15,
      "prefill_tok_s": 6.75,
      "eval_ms": 747.65,
      "eval_tokens": 8,
      "gen_tok_s": 9.36
    },
    {
      "task": 11
    },
    {
      "task": 23910
    },
    {
      "task": 47782
    },
    {
      "task": 71245
    }
  ]
}