Task 6 of 18 in ProgramBench
facebook/zstd
Astra's two runs followed the time asked: 10 min 12 s when asked for 10 min, and 2 h 32 min when asked for 2 h 30 min. Muse Wonder worked about the same whatever it was asked: 5 min 29 s when asked for 4 h 15 min, and 8 min 47 s when asked for 68 h.
- Asked for
- 10 min to 160 h
- Runs here
- 18
- 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 | 26 min 51 s | 10 min 12 s Within 5% asked 10 min | 2 h 32 min Within 5% asked 2 h 30 min |
| Sol | 2 h | 17 min 40 s | 17 min 51 s 1.78× longer asked 10 min | 2 h 31 min Within 5% asked 2 h 30 min |
| Fable | 2 h | 26 min 50 s | 8 min 50 s 1.13× shorter asked 10 min | 1 h 1 min 2.47× shorter asked 2 h 30 min |
| Opus 5.5 | 3 h | 9 h 8 min | 43 min 40 s Within 5% asked 45 min | 11 h 49 min Within 5% asked 12 h |
| Argon | 1 h 15 min | 50 min 30 s | 12 min 10 s 1.54× shorter asked 18 min 45 s | 41 min 40 s 7.20× shorter asked 5 h |
| Sol 6.1 | 2 h | 57 min 22 s | 31 min 11 s Within 5% asked 30 min | 8 h 1 min Within 5% asked 8 h |
| MiMo Pro | 40 h | 1 h 31 min | 1 h 32 min 6.55× shorter asked 10 h | 2 h 11 min 73.09× shorter asked 160 h |
| MiMo UltraSpeed | 35 h | 31 min 44 s | 52 min 41 s 9.97× shorter asked 8 h 45 min | 56 min 41 s 148.17× shorter asked 140 h |
| Muse Wonder | 17 h | 19 min 17 s | 5 min 29 s 46.53× shorter asked 4 h 15 min | 8 min 47 s 464.16× shorter asked 68 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: Yang et al., 2026