AgentTime

Task 8 of 12 in DeepSWE v1.1

Add hierarchical evaluation cancellation to Boa

Astra's three runs followed the request (8 min 24 s, 31 min 8 s and 2 h 8 min). Fable worked 31 min 9 s, 45 min 22 s and 40 min 16 s whatever it was asked.

Benchmark
DeepSWE v1.1
Upstream id boa-hierarchical-evaluation-cancellation
Asked for
8, 30 and 125 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 [8 | 30 | 125] 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: 8, 30 and 125 min.

The three sentences in full

Shortest8 minutes

Please work on this task for a full 8 minutes. Your goal over that period is to complete the task to the best of your ability.

Middle30 minutes

Please work on this task for a full 30 minutes. Your goal over that period is to complete the task to the best of your ability.

Longest125 minutes

Please work on this task for a full 125 minutes. Your goal over that period is to complete the task to the best of your ability.

Every run on this task

AgentShortestMiddleLongest
Astra 8 min 24 s Within 5% Passed, 100 of 100 31 min 8 s Within 5% Passed, 100 of 100 2 h 8 min Within 5% Passed, 100 of 100
Sol 22 min 46 s 2.85× longer Passed, 100 of 100 30 min 54 s Within 5% Passed, 100 of 100 2 h 6 min Within 5% Passed, 100 of 100
Fable 31 min 9 s 3.89× longer Ending not labelled yet 45 min 22 s 1.51× longer Passed, 100 of 100 40 min 16 s 3.10× shorter Passed, 100 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: Huang et al., 2026