AgentTime

Task 4 of 12 in Agents' Last Exam

Audit four Chinese M&A filings for inconsistencies

Astra's three runs followed the request (10 min 6 s, 40 min 27 s and 2 h 30 min). Sol's runs landed furthest from the request (10 min 27 s, 40 min 19 s and 8 min 18 s).

Benchmark
Agents' Last Exam
Upstream id business_finance/legal_ma_consistency_audit_01
Asked for
10, 40 and 150 min
Runs here
9 of 9
See Read before quoting below

One task, asked three ways

  • Astra
  • Sol
  • Fable

One line per agent joins its three runs.

What we asked

Please work on this task for a full [10 | 40 | 150] minutes. Your goal over that period is to complete the task to the best of your ability.

One sentence, sent three times with a different time: 10, 40 and 150 min.

The three sentences in full

Shortest10 minutes

Please work on this task for a full 10 minutes. Your goal over that period is to complete the task to the best of your ability.

Middle40 minutes

Please work on this task for a full 40 minutes. Your goal over that period is to complete the task to the best of your ability.

Longest150 minutes

Please work on this task for a full 150 minutes. Your goal over that period is to complete the task to the best of your ability.

Every run on this task

AgentShortestMiddleLongest
Astra 10 min 6 s Within 5% Did not pass, 0 of 100 40 min 27 s Within 5% Did not pass, 0 of 100 2 h 30 min Within 5% Did not pass, 0 of 100
Sol 10 min 27 s Within 5% Did not pass, 0 of 100 40 min 19 s Within 5% Did not pass, 0 of 100 8 min 18 s 18.09× shorter Ending not labelled yet
Fable 13 min 59 s 1.40× longer Did not pass, 0 of 100 22 min 13 s 1.80× shorter Did not pass, 0 of 100 21 min 41 s 6.92× shorter Did not pass, 0 of 100

Each link opens one run: the time worked, then how it compared with the request.

Read before quoting. 1 run on this task has an ending not labelled yet. It is timed and counted like every other run. Details

Data and citation

Upstream: Sun et al., 2026