luosuu/SWE-kokkos-bench

Coding

100 verified SWE-bench-style coding tasks from the Kokkos ecosystem.

← Back to Registry

Run this task

CLI:

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

Python:

from inspect_ai import eval
from inspect_harbor import luosuu_swe_kokkos_bench

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

Dataset information

Harbor registry luosuu/SWE-kokkos-bench
Inspect task luosuu_swe_kokkos_bench
Latest digest sha256:2d2cf7e0ad502943456bc7ea738c297f1c4b1841aff2bc967a1d1c6ab6400d68
Samples 100
Source https://github.com/Luosuu/tinker-cookbook/tree/science-rl

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