Task 18 of 18 in OSWorld
Shotcut montage project
Sol 6.1's two runs came closest to the time asked: 12 min 34 s when asked for 11 min 15 s, and 41 min 8 s when asked for 3 h. Argon worked about the same whatever it was asked: 2 h 38 min when asked for 5 min 30 s, and 2 h 28 min when asked for 1 h 28 min.
- Asked for
- 5 min 30 s to 8 h
- Runs here
- 10
- 5 agents
Every agent on this task
- Opus 5.5
- Argon
- Sol 6.1
- MiMo Pro
- MiMo UltraSpeed
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 h 35 min | 52 min 3 s | 39 min 31 s 1.66× longer asked 23 min 45 s | 53 min 42 s 7.08× shorter asked 6 h 20 min |
| Argon | 22 min | 1 h 41 min | 2 h 38 min 28.70× longer asked 5 min 30 s | 2 h 28 min 1.68× longer asked 1 h 28 min |
| Sol 6.1 | 45 min | 28 min 1 s | 12 min 34 s 1.12× longer asked 11 min 15 s | 41 min 8 s 4.38× shorter asked 3 h |
| MiMo Pro | 2 h | 3 h 43 min | 6 h 46 min 13.54× longer asked 30 min | 4 h 41 min 1.71× shorter asked 8 h |
| MiMo UltraSpeed | 1 h 20 min | 2 h 55 min | 3 h 33 min 10.65× longer asked 20 min | 2 h 45 min 1.94× shorter asked 5 h 20 min |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Yuan et al., 2026