Task 5 of 10 in TUA-Bench
Merge text document
Sol 6.1's two runs came closest to the time asked: 1 min 2 s when asked for 53 s, and 14 min 20 s when asked for 14 min. Muse Wonder's two runs landed furthest: 54 s when asked for 1 min 15 s, and 24 s when asked for 20 min.
- Asked for
- 30 s to 1 h 12 min
- Runs here
- 18
- See Read before quoting below
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 | 3 min | 6 min 57 s | 1 min 48 s 1.44× longer asked 1 min 15 s | 20 min 17 s Within 5% asked 20 min |
| Sol | 5 min | 3 min 32 s | 3 min 49 s 3.05× longer asked 1 min 15 s | 21 min 13 s 1.06× longer asked 20 min |
| Fable | 3 min | 8 min 19 s | 3 min 29 s 2.79× longer asked 1 min 15 s | 9 min 53 s 2.03× shorter asked 20 min |
| Opus 5.5 | 5 min | 5 min 5 s | 2 min 31 s 2.02× longer asked 1 min 15 s | 11 min 22 s 1.76× shorter asked 20 min |
| Argon | 2 min | 7 min 47 s | 3 min 44 s 7.48× longer asked 30 s | 9 min 58 s 1.25× longer asked 8 min |
| Sol 6.1 | 3 min 30 s | 1 min 16 s | 1 min 2 s 1.19× longer asked 53 s | 14 min 20 s Within 5% asked 14 min |
| MiMo Pro | 5 min | 1 min 13 s | 1 min 11 s 1.06× shorter asked 1 min 15 s | 3 min 47 s 5.28× shorter asked 20 min |
| MiMo UltraSpeed | 18 min | 31 s | 56 s 4.79× shorter asked 4 min 30 s | 9 min 58 s 7.23× shorter asked 1 h 12 min |
| Muse Wonder | 5 min | 1 min 8 s | 54 s 1.39× shorter asked 1 min 15 s | 24 s 50.00× shorter asked 20 min |
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.
Read before quoting. 1 run was stopped by AgentTime, so its time is a lower bound. Details
Data and citation
Upstream: Chen et al., 2026