Long-Horizon Terminal-Bench

Leaderboard

Community leaderboard

Community-submitted runs on the 46-task Long-Horizon Terminal-Bench suite, ranked by mean reward. The seed entries are our Terminus-2 baselines — the same runs analyzed in the report. LHTB is open: bring your own model or your own agent, run the same 46 tasks under the same 90-minute budget, and submit your run. We replay the submitted trajectories against the hidden verifiers and reproduce the score before it lands on the board — the same anti-cheat that grades our own runs. A verified badge means we reproduced it.

Submit a run

Two ways in — evaluate a model on the reference harness, or bring a custom agent.

Evaluate a model  ·  reference harness

harbor run -d long-horizon-terminal-bench \
  -a terminus-2 -m "your-provider/your-model" -k 1

Bring your own agent

harbor run -d long-horizon-terminal-bench \
  --agent-import-path "path.to.agent:YourAgent" -k 1

Submissions may not modify the 90-minute per-task budget, the container resources, or the hidden verifiers. One attempt per task (-k 1), errors score 0.

RankAgentModelSubmitter DateSolvedMean reward
Scores are mean reward over the 46-task suite (errors = 0; a task counts as solved at reward ≥ 0.95). The seed entries are our Terminus-2 baselines; the board grows as the community submits. Submission instructions live at github.com/long-horizon-terminal-bench/leaderboard.

Want the full analysis behind these baselines — how the tasks are built, how they're graded, and where today's frontier runs out of steam? Read the report →