Task 6 of 20 in Humanity's Last Exam
Humanity's Last Exam item 30
Opus 5.5's two runs came closest to the time asked: 4 min 44 s when asked for 4 min 15 s, and 1 h 6 min when asked for 1 h 8 min. MiMo UltraSpeed's two runs landed furthest: 5 min 31 s when asked for 8 min 45 s, and 12 min 18 s when asked for 2 h 20 min.
- Asked for
- 53 s to 2 h 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 | 17 min | 12 min 41 s | 4 min 44 s 1.11× longer asked 4 min 15 s | 1 h 6 min Within 5% asked 1 h 8 min |
| Argon | 3 min 30 s | 17 min 20 s | 1 min 22 s 1.56× longer asked 53 s | 14 min 36 s Within 5% asked 14 min |
| Sol 6.1 | 7 min | 7 min 11 s | 6 min 54 s 3.95× longer asked 1 min 45 s | 28 min 21 s Within 5% asked 28 min |
| MiMo Pro | 30 min | 1 min 16 s | 58 min 19 s 7.77× longer asked 7 min 30 s | 1 h 8 min 1.76× shorter asked 2 h |
| MiMo UltraSpeed | 35 min | 17 s | 5 min 31 s 1.58× shorter asked 8 min 45 s | 12 min 18 s 11.39× shorter asked 2 h 20 min |
| Muse Wonder | 22 min | 9 min 9 s | 6 min 27 s 1.17× longer asked 5 min 30 s | 15 min 22 s 5.73× shorter asked 1 h 28 min |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Phan et al., 2025