terminal-bench-pro/terminal-bench-pro

Coding

Terminal-Bench Pro: tasks across 8 domains — data processing, games, debugging, sysadmin, scientific computing, SWE, ML, and security — extending Terminal-Bench with harder real-world scenarios.

← Back to Registry

Run this task

CLI:

inspect eval inspect_harbor/terminal_bench_pro --model openai/gpt-5

Python:

from inspect_ai import eval
from inspect_harbor import terminal_bench_pro

eval(terminal_bench_pro(), model="openai/gpt-5")

Dataset information

Harbor registry	terminal-bench-pro/terminal-bench-pro
Inspect task	`terminal_bench_pro`
Latest digest	sha256:1e0df73902c63a708b75e5a4b633eaa21524583904fdadaf937203280f037461
Samples	200
Paper	arxiv
Source	https://github.com/alibaba/terminal-bench-pro

See Task Parameters for the parameter set shared across all Harbor tasks.