feat: add Terminal-Bench 4.0 as a separate Modal-backed benchmark - #72

Open

abhinav-pola wants to merge 6 commits into

mainfrom

devin/1788466045-terminal-bench-4

Commits on Sep 3, 2026

feat: add Terminal-Bench 4.0 as a separate Modal-backed benchmark

devin-ai-integration[bot]
and
abhinav-pola
committed
b372806
feat: resolve terminal-bench-4 images by Modal image id and harden sandbox lifetimes

devin-ai-integration[bot]
and
abhinav-pola
committed
fc97db9
fix: add TB4 scorer, full Modal image map, artifact path normalization, and root-user deviation metadata

devin-ai-integration[bot]
and
abhinav-pola
committed
0797bff
fix: install agent CLI prerequisites via apt-get, dnf or apk so Fedora task images build

devin-ai-integration[bot]
and
abhinav-pola
committed
7481861
docs: note portable package install for the Fedora task image

devin-ai-integration[bot]
and
abhinav-pola
committed
2480a3d
fix: install bash on apk-based agent images

devin-ai-integration[bot]
and
abhinav-pola
committed
0712971