Task 7 of 8 in AssistantBench

AssistantBench validation task 18

Muse Wonder's two runs came closest to the time asked: 10 min 30 s when asked for 5 min 30 s, and 1 h 29 min when asked for 1 h 28 min. Fable worked about the same whatever it was asked: 40 min when asked for 1 min 15 s, and 40 min when asked for 20 min.

Benchmark
AssistantBench
Upstream id validation:32
Asked for
1 min 15 s to 2 h
Runs here
18
9 agents

Every agent on this task

  • Astra
  • Sol
  • Fable
  • Opus 5.5
  • Argon
  • Sol 6.1
  • MiMo Pro
  • MiMo UltraSpeed
  • Muse Wonder

One line per agent joins its short and long runs.

What we asked

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

Each agent read it twice, with its own short and long ask, as listed below.

Every run on this task

Data and citation

Upstream: Yoran et al., 2024 (EMNLP)