The Prompting Company

Benchmark report

Task verified

3m 5s elapsed

The saved attempt met the documented success criterion.

Experience score

94/ 100

Agent-Ready
  • AInstall
  • SAuth
  • SExecute
  • ADocs
  • SCurrent
Experience measures friction in the tested task. The verdict says whether it succeeded.
Agent profile for this run: Install A, Auth S, Execute S, Docs A, Current SAInstallSAuthSExecuteADocsSCurrent
Agent profile · this run
Judge's note · what slowed the agent most
Two small detours: pip's PEP 668 block requiring a venv, and the docs HTML page dumping 106KB of markup before the .md variant gave readable content.

Where the time went

3:06 agent time
  • Research0:13
  • Attempt1:55
  • Verify0:58
3m 5sTotal elapsed
1:06Time to first success
0:00Waited on the owner

Judge reason

Saved code parallel-search/search.py calls the documented Parallel Search API via the official parallel-web 1.3.5 SDK (client.search(objective=..., search_queries=[...])) with the exact query 'PostgreSQL transaction isolation levels'. run_output.txt shows a live run: search_id search_9ad09392ef7615f24804277151468971, result_count 10, first_result_title 'Documentation: 18: 13.2. Transaction Isolation', first_result_url https://www.postgresql.org/docs/current/transaction-iso.html — all matching results[0] in response.json (keys results/search_id/session_id/usage/warnings/metadata; 10 results, all absolute https URLs and nonempty titles). An independent raw-HTTP run (search_curl.sh) corroborates liveness: curl_status.txt http_status=200, curl_response_headers.txt 'HTTP/2 200, server: Parallel, x-request-id: 1acfed0dc2bbc9b60e704832efcca72e' matching curl_response.json search_id search_1acfed0dc2bbc9b60e704832efcca72e (a distinct call from the SDK one). Key was obtained via benchmark-request (outcome 'provided') and read from a private file, never printed. No mocking or hard-coded responses.

Attempt record

  1. 1:02Answered from saved key: Parallel API keyno wait
  2. —First success verified

Attempt artifacts saved.