Task 9 of 18 in ProgramBench
doxygen/doxygen
Sol 6.1's two runs followed the time asked: 26 min 38 s when asked for 26 min 15 s, and 7 h when asked for 7 h. Muse Wonder's one run landed furthest: 2 h 12 min when asked for 72 h.
- Asked for
- 26 min 15 s to 200 h
- Runs here
- 11
- 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 | 4 h | 19 h 4 min | 21 min 43 s 2.76× shorter asked 1 h | 14 h 54 min 1.07× shorter asked 16 h |
| Argon | 1 h 45 min | 42 min 55 s | 28 min 31 s 1.09× longer asked 26 min 15 s | 2 h 21 min 2.98× shorter asked 7 h |
| Sol 6.1 | 1 h 45 min | 56 min 9 s | 26 min 38 s Within 5% asked 26 min 15 s | 7 h Within 5% asked 7 h |
| MiMo Pro | 50 h | 43 min 55 s | 2 h 43 min 4.60× shorter asked 12 h 30 min | 1 h 10 min 172.34× shorter asked 200 h |
| MiMo UltraSpeed | 15 h | 24 min 5 s | 57 min 58 s 3.88× shorter asked 3 h 45 min | 1 h 14 min 48.89× shorter asked 60 h |
| Muse Wonder | None | None | No run | 2 h 12 min 32.63× shorter asked 72 h |
No runs match these filters.
Runtime is each agent’s run with no time asked.
Data and citation
Upstream: Yang et al., 2026