Benchmark report

Task verified

9m 34s elapsed

The saved attempt met the documented success criterion.

Experience score

100/ 100

Agent-Ready
  • SInstall
  • SAuth
  • SExecute
  • SDocs
  • SCurrent
Experience measures friction in the tested task. The verdict says whether it succeeded.
Agent profile for this run: Install S, Auth S, Execute S, Docs S, Current SSInstallSAuthSExecuteSDocsSCurrent
Agent profile · this run
Judge's note · what slowed the agent most
Essentially nothing: doc fetch, key request, one script run, success in ~50s with no retries or errors.

Where the time went

9:34 agent time
  • Starting0:36
  • Research0:27
  • Attempt5:32
  • Verify2:59
9m 34sTotal elapsed
0:50Time to first success
0:00Waited on the owner

Judge reason

Saved evidence proves a single live authenticated Tavily call met the frozen task. request.json/search_mvcc_vacuum.py show POST https://api.tavily.com/search with Bearer auth and body {query: "MVCC vacuum", include_domains: ["postgresql.org"], include_domains_mode: "restrict", exclude_domains: ["wiki.postgresql.org"]} - no site: operator or domain words in the query. run_output.txt shows HTTP status 200 and 10 titles+URLs printed; response.json is a genuine Tavily payload (request_id ec43e171-65c3-4f5c-b7f4-2d176f893ef2, response_time 3.2, per-result scores and content). I re-parsed response.json independently: 10 results, every host www.postgresql.org, none wiki.postgresql.org, >=3 results. Filtering was server-side; the script's host check runs after the fact on the full returned set and filters nothing. Fetched docs confirm include_domains/exclude_domains/include_domains_mode are GA (only include_published_date is marked beta), so no beta-stop applied. No mocking or hard-coded data.

Attempt record

  1. 5:14Answered from saved key: Tavily API keyno wait
  2. —First success verified

Attempt artifacts saved.