A Humanity's Last Exam task from the paper, not in v1.1

Humanity's Last Exam item 9

Astra's two runs came closest to the time asked: 1 min 24 s when asked for 1 min 15 s, and 20 min 13 s when asked for 20 min. Fable worked about the same whatever it was asked: 4 min 31 s when asked for 1 min 15 s, and 4 min 37 s when asked for 20 min.

Benchmark
Humanity's Last Exam
Upstream id 673818e39b3842b34824240d
Asked for
1 min 15 s to 20 min
Runs here
6
3 agents

Every agent on this task

  • Astra
  • Sol
  • Fable

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

Every run on Humanity's Last Exam item 9, one row per agent
AgentForecastRuntimeShort askLong ask
Astra5 min3 min 13 s 1 min 24 s 1.12× longer asked 1 min 15 s 20 min 13 s Within 5% asked 20 min
Sol5 min2 min 29 s 2 min 31 s 2.01× longer asked 1 min 15 s 20 min 17 s Within 5% asked 20 min
Fable10 min3 min 34 s 4 min 31 s 3.61× longer asked 1 min 15 s 4 min 37 s 4.33× shorter asked 20 min

Runtime is the paper’s estimate from each agent’s timed runs.

Data and citation

Upstream: Phan et al., 2025