Task 6 of 8 in AssistantBench
AssistantBench validation task 16
Argon's two runs came closest to the time asked: 1 min 2 s when asked for 53 s, and 14 min 39 s when asked for 14 min. Astra worked about the same whatever it was asked: 2 min 58 s when asked for 1 min 15 s, and 4 min 52 s when asked for 20 min.
- Asked for
- 53 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
No runs match these filters.
Runtime is a v1.1 agent’s run with no time asked, and for the paper’s three agents the paper’s estimate from their timed runs.
Data and citation
Upstream: Yoran et al., 2024 (EMNLP)