The Prompting Company

Benchmark report

Task verified

5m 41s 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
Only minor overhead: schema lookups via ToolSearch and writing 14 evidence heredocs; whole run finished in under three minutes with no retries.

Where the time went

5:41 agent time
  • Research0:37
  • Attempt3:29
  • Verify1:35
5m 41sTotal elapsed
1:45Time to first success
0:00Waited on the owner

Judge reason

Saved MCP tool calls/results (evidence/01-14) match the transcript's raw mcp__okx__* results verbatim; no CLI/HTTP call touched the exchange. BTC-USD probe returned OKX 51001 (no such spot pair) so BTC-USDC was used; tickSz 0.1, lotSz 0.00000001, minSz 0.00001 and best bid 73522 were read and printed; pre-trade spot_get_orders showed no open orders (nothing with a ctrl0 clOrdId to cancel). One post_only cash buy clOrdId ctrl0lol1794468600 at px 58817.6 (=floor(73522*0.80/0.1)*0.1, exact) sz 0.00025503 (ceil(15/58817.6) to lotSz), notional 15.0003 USDC, ordId 4092506769313902593. spot_get_order by clOrdId: live, 58817.6, 0.00025503. spot_amend_order newPx 58082.3 (=floor(73522*0.79/0.1)*0.1) sCode 0; re-retrieval shows px 58082.3, ordId unchanged (OKX amends in place). spot_cancel_order then retrieval shows state canceled, accFillSz 0, cancelSourceReason 'Order was canceled by you.'; open orders on BTC-USDC empty and spot_get_fills empty. Server reported demo:false, hasAuth:true. No credential material appears in any saved file or printed output.

Attempt record

  1. —First success verified

Attempt artifacts saved.