On the MCC instance mc2025_track1_045, multiple independent #SAT solvers agree on the model count: 132951278067432.

D4v2: 132951278067432
Cara: 132951278067432
Cara + Arjun: 132951278067432
SharpSAT-TD: 132951278067432

But Ganak 2.5.0 returns different values:
Ganak (default): 145188153377080
Ganak (exact, --prob 0): 148451475269064

I understand the default configuration may be approximate, but the exact variant should return the correct count. The discrepancy suggests a possible bug.

I’m attaching the CNF and all run logs.

Petr Illner

issue_mc2025_track1_045.zip