Task 10 of 12 in TUA-Bench

EPW parquet check

Astra's three runs landed closest to the request (3 min 42 s, 6 min 23 s and 25 min 23 s). Fable worked 7 min 17 s, 9 min 23 s and 12 min 16 s whatever it was asked.

Benchmark: TUA-Bench; Upstream id 011-epw-parquet-check
Asked for: 1.5, 6 and 25 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 [1.5 | 6 | 25] 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: 1.5, 6 and 25 min.

The three sentences in full

Shortest1.5 minutes

Please work on this task for a full 1.5 minutes. Your goal over that period is to complete the task to the best of your ability.

Middle6 minutes

Please work on this task for a full 6 minutes. Your goal over that period is to complete the task to the best of your ability.

Longest25 minutes

Please work on this task for a full 25 minutes. Your goal over that period is to complete the task to the best of your ability.

Every run on this task

Every run on EPW parquet check, one row per agent
Agent	Shortest	Middle	Longest
Astra	3 min 42 s 2.47× longer Did not pass, 66 of 100	6 min 23 s 1.06× longer Ending not labelled yet	25 min 23 s Within 5% Did not pass, 66 of 100
Sol	4 min 19 s 2.88× longer Did not pass, 66 of 100	6 min 52 s 1.15× longer Ending not labelled yet	27 min 6 s 1.08× longer Did not pass, 66 of 100
Fable	7 min 17 s 4.86× longer Did not pass, 66 of 100	9 min 23 s 1.56× longer Did not pass, 66 of 100	12 min 16 s 2.04× shorter Did not pass, 66 of 100

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

Read before quoting. 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: Chen et al., 2026