Python

MolBench

MolBench — a molecular-visualization benchmark & harness for CI (LLM) assistants: natural language → Mol*/ipymolstar API calls, graded deterministically, plus VLM-judged visual tasks.

PythonEmerging

GitHub

Stars

—

Forks

—

Contributors

Last push

10d ago

Recent commits

Latest commits.

docs(readme): add a worked example task near the top
7c5ef50Dan Bolser10d ago
Tiered grading: full 7-model results (rescue-only policy)
15b5acaDan Bolser17d ago
Add 'Tiered grading (preliminary)' section to the leaderboard page
fea2ca4Dan Bolser17d ago
Redesign Pages site: professional leaderboard + 9-row evaluator gallery
2493668Dan Bolser17d ago
Wire the escalating cascade into the runner (--escalate / --judge)
4f60e05Dan Bolser17d ago

Top contributors

Builders behind this project.

dbolser

1 commits