AgentTime

Task 1 of 3 in PaperBench

Bridging Data Gaps

Astra's three runs followed the request (12 min 39 s, 50 min 38 s and 3 h 21 min). Fable's runs landed furthest from the request (5 min 18 s, 15 min 13 s and 45 min 29 s).

Benchmark
PaperBench
Asked for
13, 50 and 200 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 [12.5 | 50 | 200] 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: 13, 50 and 200 min.

The three sentences in full

Shortest12.5 minutes

Please work on this task for a full 12.5 minutes. Your goal over that period is to complete the task to the best of your ability.

Middle50 minutes

Please work on this task for a full 50 minutes. Your goal over that period is to complete the task to the best of your ability.

Longest200 minutes

Please work on this task for a full 200 minutes. Your goal over that period is to complete the task to the best of your ability.

Every run on this task

AgentShortestMiddleLongest
Astra 12 min 39 s Within 5% Not graded 50 min 38 s Within 5% Not graded 3 h 21 min Within 5% Not graded
Sol 6 min 11 s 2.02× shorter Score 7.3 of 100 51 min 15 s Within 5% Score 23.8 of 100 3 h 20 min Within 5% Score 34.4 of 100
Fable 5 min 18 s 2.36× shorter Ending not labelled yet 15 min 13 s 3.29× shorter Ending not labelled yet 45 min 29 s 4.40× shorter Stopped by an error

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

Read before quoting: 2 notes
  1. 1 run was stopped by AgentTime, so its time is a lower bound. Details
  2. 2 runs on this task have an ending not labelled yet. They are timed and counted like every other run. Details

Data and citation

Upstream: Starace et al., 2025