Results for “agentbench”

Execute the full AgentHub competition lifecycle in a single command: initialize, capture baseline, spawn agents, evaluate results, and merge the winner.

Spawns multiple parallel AI agents that compete on the same task using isolated git worktrees, evaluates results, and merges the best solution.