Code Generation on LiveCodeBench 50/50 calibration (test)
64.9CR (%)Diagnostic learned router
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Diagnostic learned router2026.05 | 64.9 | 36.5 | |
| Optimal pair2026.05 | 56.1 | 32.9 | |
| Optimal subsequence2026.05 | 51.2 | 31.5 | |
| Full fixed chain2026.05 | 41.4 | 27.7 | |
| Always-expensive2026.05 | 0 | — |