ai-forever/harness-bench-fast
Coding
Self-contained file-operation agent benchmark.
Run this task
CLI:
inspect eval inspect_harbor/ai_forever_harness_bench_fast --model openai/gpt-5Python:
from inspect_ai import eval
from inspect_harbor import ai_forever_harness_bench_fast
eval(ai_forever_harness_bench_fast(), model="openai/gpt-5")Dataset information
| Harbor registry | ai-forever/harness-bench-fast |
| Inspect task | ai_forever_harness_bench_fast |
| Latest digest | sha256:c8376d1cd706a6325a7b7c8dbac360e59a5c9645d3075dcc83381b1c333899c0 |
| Samples | 231 |
| Source | https://github.com/ai-forever/harness-bench-fast |
See Task Parameters for the parameter set shared across all Harbor tasks.