AgentTime

Task 11 of 12 in TUA-Bench

Fix MP3 metadata

Astra's three runs landed closest to the request (1 min 32 s, 5 min 24 s and 20 min 16 s). Fable's runs landed furthest from the request (45 s, 2 min 57 s and 1 min 49 s).

Benchmark
TUA-Bench
Upstream id 079-fix-mp3-metadata
Asked for
1.25, 5 and 20 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 [1.25 | 5 | 20] 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: 1.25, 5 and 20 min.

The three sentences in full

Shortest1.25 minutes

Please work on this task for a full 1.25 minutes. Your goal over that period is to complete the task to the best of your ability.

Middle5 minutes

Please work on this task for a full 5 minutes. Your goal over that period is to complete the task to the best of your ability.

Longest20 minutes

Please work on this task for a full 20 minutes. Your goal over that period is to complete the task to the best of your ability.

Every run on this task

AgentShortestMiddleLongest
Astra 1 min 32 s 1.23× longer Passed, 100 of 100 5 min 24 s 1.08× longer Ending not labelled yet 20 min 16 s Within 5% Passed, 100 of 100
Sol 1 min 57 s 1.55× longer Not graded 5 min 45 s 1.15× longer Passed, 100 of 100 21 min 6 s 1.05× longer Passed, 100 of 100
Fable 45 s 1.68× shorter Stopped by an error 2 min 57 s 1.69× shorter Stopped by an error 1 min 49 s 10.98× 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. 3 runs were stopped by AgentTime, so their time is a lower bound. Details
  2. 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: Chen et al., 2026