{
  "test_date": "2026-10-07",
  "reference_implementation": {
    "repository": "https://github.com/browser-use/jev-ultrafast",
    "commit": "1231850a0bf1a0c0341fe408ef1668dbbfdfac46"
  },
  "models": {
    "decision_model": "jev-1.13.0",
    "text_helper": "opencode-go/deepseek-v4.1-flash"
  },
  "task": "Open the exact Bodum Bistro electric coffee grinder product detail page from Imerco's coffee-grinder category.",
  "successful_runs": [
    {
      "run": 7,
      "agent_elapsed_ms": 11503,
      "independent_verification_wait_ms": 472,
      "recording_elapsed_ms": 11975,
      "jev_decisions": 5,
      "browser_actions": 3,
      "median_jev_decision_ms": 290,
      "jev_input_tokens": 27163,
      "estimated_jev_cost_usd": 0.001141,
      "text_helper_latency_ms": 9151,
      "text_helper_input_tokens": 16766,
      "text_helper_output_tokens": 28,
      "outcome_checks_passed": 5,
      "outcome_checks_total": 5
    },
    {
      "run": 6,
      "agent_elapsed_ms": 27084,
      "independent_verification_wait_ms": 0,
      "recording_elapsed_ms": 27085,
      "jev_decisions": 6,
      "browser_actions": 3,
      "median_jev_decision_ms": 307,
      "jev_input_tokens": 34322,
      "estimated_jev_cost_usd": 0.001442,
      "text_helper_latency_ms": 24104,
      "text_helper_input_tokens": 16647,
      "text_helper_output_tokens": 28,
      "outcome_checks_passed": 5,
      "outcome_checks_total": 5
    }
  ],
  "cold_start_failure": {
    "run": 2,
    "elapsed_ms": 45814,
    "jev_decisions": 120,
    "browser_actions": 0,
    "jev_input_tokens": 702709,
    "estimated_jev_cost_usd": 0.029514,
    "failure": "The cookie modal covered the search field. The execution guard rejected the covered target, but the failed choice was not retained in history, so Jev selected it repeatedly until the model-call budget was exhausted."
  },
  "test_boundaries": {
    "isolated_browser_profile": true,
    "cookie_consent_preconditioned_for_successful_runs": true,
    "cart_changed": false,
    "checkout_started": false,
    "purchase_attempted": false,
    "account_used": false,
    "video_speed": "1x"
  }
}