Task 13 of 18 in OSWorld
Calc macros, contracts, and emails
Astra's two runs came closest to the time asked: 21 min 9 s when asked for 15 min, and 4 h 11 min when asked for 4 h 10 min. MiMo Pro's two runs landed furthest: 2 h 28 min when asked for 13 min 45 s, and 6 h 41 min when asked for 3 h 40 min.
- Asked for
- 5 min to 6 h
- Runs here
- 16
- 8 agents
Every agent on this task
- Astra
- Sol
- Fable
- 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 |
|---|---|---|---|---|
| Astra | 1 h | 43 min 55 s | 21 min 9 s 1.41× longer asked 15 min | 4 h 11 min Within 5% asked 4 h 10 min |
| Sol | 45 min | 54 min 46 s | 34 min 45 s 2.32× longer asked 15 min | 53 min 9 s 4.70× shorter asked 4 h 10 min |
| Fable | 36 min | 45 min 2 s | 26 min 58 s 1.80× longer asked 15 min | 1 h 28 min 2.85× shorter asked 4 h 10 min |
| Opus 5.5 | 45 min | 1 h 14 min | 48 min 22 s 4.30× longer asked 11 min 15 s | 2 h 16 min 1.32× shorter asked 3 h |
| Argon | 20 min | 47 min 34 s | 15 min 34 s 3.11× longer asked 5 min | 34 min 47 s 2.30× shorter asked 1 h 20 min |
| Sol 6.1 | 1 h | 18 min 58 s | 15 min 35 s Within 5% asked 15 min | 1 h 3.97× shorter asked 4 h |
| MiMo Pro | 55 min | 3 h 54 min | 2 h 28 min 10.73× longer asked 13 min 45 s | 6 h 41 min 1.82× longer asked 3 h 40 min |
| MiMo UltraSpeed | 1 h 30 min | 6 h 40 min | 3 h 19 min 8.86× longer asked 22 min 30 s | 6 h 31 min 1.09× longer asked 6 h |
No runs match these filters.
Runtime is a v1.1 agent’s run with no time asked, and for the paper’s three agents the paper’s estimate from their timed runs.
Data and citation
Upstream: Yuan et al., 2026