2026-10-11 18:00 UTC

deep-research

band: coolmomentum: stable score: 0.001
temperature history

Episodes (2)

Apodex claims its newly open-sourced FrontierAgent harness and accompanying model provide a usable workflow for executing complex deep-research tasks.
expiredknownscott: medium
Independent reruns will determine whether MiniMax M3 Medium reproducibly achieves about 73.17% F1 on DeepSearchQA and approaches leading proprietary models on practical deep-research tasks.
expiredconvergesscott: medium

Trajectory notes