terminal-bench/terminal-bench-cpu-only
Coding
Assistants
Terminal-Bench (CPU-only): the current Terminal-Bench task set minus the 3 tasks that require an H100 GPU, so the full suite runs on CPU-only sandboxes.
Run this task
CLI:
inspect eval inspect_harbor/terminal_bench_cpu_only --model openai/gpt-5Python:
from inspect_ai import eval
from inspect_harbor import terminal_bench_cpu_only
eval(terminal_bench_cpu_only(), model="openai/gpt-5")Dataset information
| Harbor registry | terminal-bench/terminal-bench-cpu-only |
| Inspect task | terminal_bench_cpu_only |
| Latest digest | sha256:b440941ff70a00335fa906e3b1a5407c3e72c20e975fa4d7d61b803bf80dd5b5 |
| Samples | 63 |
| Paper | arxiv |
| Source | https://github.com/harbor-framework/terminal-bench |
See Task Parameters for the parameter set shared across all Harbor tasks.