STEMGym Benchmark

What is STEMGym?

STEMGym is an open-source Gymnasium benchmark of 15 physics-simulated STEM worlds across five materials and four characterization tasks, scored by DEC-AUC.

Released

2026-06-28

Evaluates

General AI, Language & Knowledge, cs.LG

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.29592

STEMGym paper

• Paper https://arxiv.org/abs/2606.29592

Which sources cite STEMGym?

• STEMGym

Related General AI, Language & Knowledge, cs.LG benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A3S-Bench
• AARRI-Bench
• ABC-Bench
• ActBench