Task 9 of 20 in Humanity's Last Exam
Humanity's Last Exam item 33
Sol 6.1's two runs came closest to the time asked: 5 min 39 s when asked for 2 min 30 s, and 40 min 12 s when asked for 40 min. MiMo UltraSpeed worked about the same whatever it was asked: 15 min 25 s when asked for 30 min, and 10 min 47 s when asked for 8 h.
- Asked for
- 53 s to 8 h
- 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 | 14 min | 16 min 11 s | 18 min 11 s 5.19× longer asked 3 min 30 s | 1 h 8 min 1.21× longer asked 56 min |
| Argon | 3 min 30 s | 7 min 59 s | 3 min 40 s 4.18× longer asked 53 s | 13 min 33 s Within 5% asked 14 min |
| Sol 6.1 | 10 min | 11 min 32 s | 5 min 39 s 2.26× longer asked 2 min 30 s | 40 min 12 s Within 5% asked 40 min |
| MiMo Pro | 40 min | 29 min 23 s | 1 h 17 min 7.71× longer asked 10 min | 2 h 12 min 1.21× shorter asked 2 h 40 min |
| MiMo UltraSpeed | 2 h | 7 min 46 s | 15 min 25 s 1.95× shorter asked 30 min | 10 min 47 s 44.52× shorter asked 8 h |
| Muse Wonder | 9 min | 1 h 12 min | 8 min 1 s 3.57× longer asked 2 min 15 s | 1 h 25 min 2.36× longer asked 36 min |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Phan et al., 2025