Task 11 of 20 in Humanity's Last Exam
Humanity's Last Exam item 35
Sol 6.1's two runs came closest to the time asked: 23 s when asked for 8 s, and 2 min 9 s when asked for 2 min. MiMo UltraSpeed worked about the same whatever it was asked: 13 s when asked for 1 min 15 s, and 22 s when asked for 20 min.
- Asked for
- 8 s to 20 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 | 1 min | 25 s | 52 s 3.49× longer asked 15 s | 4 min 44 s 1.18× longer asked 4 min |
| Argon | 30 s | 25 s | 36 s 4.84× longer asked 8 s | 2 min 23 s 1.19× longer asked 2 min |
| Sol 6.1 | 30 s | 21 s | 23 s 3.05× longer asked 8 s | 2 min 9 s 1.07× longer asked 2 min |
| MiMo Pro | 2 min | 13 min 18 s | 10 min 20.01× longer asked 30 s | 1 min 47 s 4.49× shorter asked 8 min |
| MiMo UltraSpeed | 5 min | 25 s | 13 s 5.86× shorter asked 1 min 15 s | 22 s 55.05× shorter asked 20 min |
| Muse Wonder | 30 s | 7 s | 17 s 2.25× longer asked 8 s | 1 min 1 s 1.98× shorter asked 2 min |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Phan et al., 2025