3-run cache timing evidence: test_character_creation_three_flows.py on feat/llm-response-cache-2026-05-16 @ 6ab6563

cache_timing_3run_evidence.json

{
  "benchmark": "test_character_creation_three_flows 3-run cache timing",
  "git_head": "6ab6563f86c5b609d9299c34f6e48d45f5997ed9",
  "branch": "feat/llm-response-cache-2026-05-16",
  "runs": [
    {
      "run": "cold",
      "elapsed_seconds": 484.0,
      "cache_entries_after": 34,
      "exit_code": 1,
      "hits": 4,
      "misses": 15,
      "hit_rate_pct": 21.1
    },
    {
      "run": "warm",
      "elapsed_seconds": 82.6,
      "cache_entries_after": 38,
      "exit_code": 1,
      "hits": 21,
      "misses": 1,
      "hit_rate_pct": 95.5
    },
    {
      "run": "hot",
      "elapsed_seconds": 88.3,
      "cache_entries_after": 38,
      "exit_code": 1,
      "hits": 23,
      "misses": 0,
      "hit_rate_pct": 100.0
    }
  ],
  "speedup_warm_vs_cold": 5.9,
  "speedup_hot_vs_cold": 5.5,
  "gemini_call_reduction_pct": 81.8
}