Coding & Software
Decompile-Bench
Two million binary-source function pairs for real-world LLM binary decompilation.
19,327items
2subjects
100%observed
Modelsubject type
Response matrix
Every model, scored item by item.
Each row is an AI model and each column an item, ordered so the strongest models and easiest items gather toward one corner. 2 subjects × 19,327 items, 100% of cells evaluated.
Fit to width. Hover for subject & item; click a cell for details.
lowhighUnobserved
Scale: 0 to 1 (per split): fraction of decompilation attempts that produced a valid, recompilable output — a pass rate, not a single binary outcome.
Sample items
What the questions look like — and how subjects answer.
A spread of items across the difficulty range, each shown with a few subjects' actual answers.
Subjects
The models, agents, and reward models evaluated.
2 subjects, ranked by mean response across this benchmark's items.
- 1e9d07cea0.991
- 259a2ea360.966