Task 2 of 20 in Humanity's Last Exam
Humanity's Last Exam item 26
Sol 6.1's two runs came closest to the time asked: 3 min 20 s when asked for 1 min 30 s, and 24 min 20 s when asked for 24 min. Opus 5.5 worked about the same whatever it was asked: 10 min 12 s when asked for 3 min 15 s, and 10 min 52 s when asked for 52 min.
- Asked for
- 30 s to 1 h 40 min
- Runs here
- 12
- 6 agents
Every agent on this task
- 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
| Agent | Forecast | Runtime | Short ask | Long ask |
|---|---|---|---|---|
| Opus 5.5 | 13 min | 6 min 59 s | 10 min 12 s 3.14× longer asked 3 min 15 s | 10 min 52 s 4.79× shorter asked 52 min |
| Argon | 2 min | 6 min 35 s | 4 min 57 s 9.90× longer asked 30 s | 8 min 35 s 1.07× longer asked 8 min |
| Sol 6.1 | 6 min | 3 min 25 s | 3 min 20 s 2.22× longer asked 1 min 30 s | 24 min 20 s Within 5% asked 24 min |
| MiMo Pro | 25 min | 23 min 48 s | 54 min 12 s 8.67× longer asked 6 min 15 s | 1 h 23 min 1.20× shorter asked 1 h 40 min |
| MiMo UltraSpeed | 20 min | 2 min 59 s | 5 min 13 s Within 5% asked 5 min | 11 min 30 s 6.96× shorter asked 1 h 20 min |
| Muse Wonder | 6 min | 17 min 31 s | 5 min 13 s 3.47× longer asked 1 min 30 s | 9 min 33 s 2.51× shorter asked 24 min |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Phan et al., 2025