MDArena, a 50-task benchmark from active biomolecular simulation projects, shows the strongest evaluated agent, Codex GPT-5.5 XHigh, solves 24/50 tasks (48%) at strict pass, with hard workflows nearly unsolved.
BMC Research Notes , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.chem-ph 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows
MDArena, a 50-task benchmark from active biomolecular simulation projects, shows the strongest evaluated agent, Codex GPT-5.5 XHigh, solves 24/50 tasks (48%) at strict pass, with hard workflows nearly unsolved.