Task 7 of 18 in ProgramBench
lfos/calcurse
Astra's two runs followed the time asked: 25 min 14 s when asked for 25 min, and 6 h 41 min when asked for 6 h 40 min. MiMo Pro worked about the same whatever it was asked: 53 min 52 s when asked for 3 h 45 min, and 1 h 31 min when asked for 60 h.
- Asked for
- 22 min 30 s to 60 h
- Runs here
- 17
- 9 agents
Every agent on this task
- Astra
- Sol
- Fable
- 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 |
|---|---|---|---|---|
| Astra | 2 h | 1 h 8 min | 25 min 14 s Within 5% asked 25 min | 6 h 41 min Within 5% asked 6 h 40 min |
| Sol | 2 h | 10 min 11 s | 23 min 42 s 1.05× shorter asked 25 min | 13 min 54 s 28.79× shorter asked 6 h 40 min |
| Fable | 2 h | 52 min 56 s | 8 min 41 s 2.88× shorter asked 25 min | 5 h 23 min 1.24× shorter asked 6 h 40 min |
| Opus 5.5 | 2 h 45 min | 19 h 47 min | 9 min 1 s 4.57× shorter asked 41 min 15 s | 10 h 35 min Within 5% asked 11 h |
| Argon | 1 h 30 min | 1 h 46 min | 43 min 13 s 1.92× longer asked 22 min 30 s | 1 h 34 min 3.82× shorter asked 6 h |
| Sol 6.1 | 1 h 30 min | 48 min 6 s | 23 min 3 s Within 5% asked 22 min 30 s | 6 h Within 5% asked 6 h |
| MiMo Pro | 15 h | 1 h 50 min | 53 min 52 s 4.18× shorter asked 3 h 45 min | 1 h 31 min 39.58× shorter asked 60 h |
| MiMo UltraSpeed | 9 h | 49 min 35 s | 1 h 21 min 1.66× shorter asked 2 h 15 min | 1 h 43 min 21.06× shorter asked 36 h |
| Muse Wonder | 30 h | 33 min 49 s | 42 min 59 s 10.47× shorter asked 7 h 30 min | No run |
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: Yang et al., 2026