A Humanity's Last Exam task from the paper, not in v1.1
Humanity's Last Exam item 17
Sol's two runs came closest to the time asked: 1 min 27 s when asked for 1 min 15 s, and 20 min 18 s when asked for 20 min. Fable worked about the same whatever it was asked: 2 min 17 s when asked for 1 min 15 s, and 2 min 22 s when asked for 20 min.
- Asked for
- 1 min 15 s to 20 min
- Runs here
- 6
- 3 agents
Every agent on this task
- Astra
- Sol
- Fable
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
| Agent | Forecast | Runtime | Short ask | Long ask |
|---|---|---|---|---|
| Astra | 5 min | 3 min 29 s | 1 min 29 s 1.19× longer asked 1 min 15 s | 20 min 27 s Within 5% asked 20 min |
| Sol | 5 min | 56 s | 1 min 27 s 1.17× longer asked 1 min 15 s | 20 min 18 s Within 5% asked 20 min |
| Fable | 3 min | 2 min 2 s | 2 min 17 s 1.83× longer asked 1 min 15 s | 2 min 22 s 8.46× shorter asked 20 min |
No runs match these filters.
Runtime is the paper’s estimate from each agent’s timed runs.
Data and citation
Upstream: Phan et al., 2025