Solved
15,183
81.0% coverage
Research in progress
Observatory is preparing and verifying real benchmark data. Public numbers will be released only after that verification is complete. Placeholder and test results stay hidden until then.
Mathematical optimization
Linear relaxations, branch-and-bound, cutting planes, and mixed discrete-continuous models.
Domain-native results grouped by model class
| model class | Family | Solved | Coverage | Δ solved | Median | Best baseline | Status |
|---|---|---|---|---|---|---|---|
| MIPLIB | General mixed-integer models | 4,256of 5,600 | 76.0% | +69 | 15.80 s | SCIP | stable |
| 0–1 ILP | Binary decision models | 3,546of 4,300 | 82.5% | +74 | 5.24 s | HiGHS | stable |
| Network flow | Structured linear networks | 2,907of 3,150 | 92.3% | +18 | 810 ms | HiGHS | stable |
| Scheduling | Time-indexed MILP | 2,406of 3,000 | 80.2% | +47 | 12.60 s | SCIP | experimental |
| Packing | Knapsack and placement | 2,068of 2,700 | 76.6% | +36 | 9.43 s | CBC | research |
Coverage on comparable experiments only
What moved this domain
Presolve and network detection close most of the structured-model gap.
Domain-native outcome classes
What makes this view reproducible