This model seems to be solving just fine:
BLOCKS OF EQUATIONS 81 SINGLE EQUATIONS 81
Gurobi 24.4.1 r50296 Released Dec 20, 2014 WEI x86 64bit/MS Windows
Gurobi full license.
Root relaxation: objective 0.000000e+00, 41 iterations, 0.00 seconds
Nodes | Current Node | Objective Bounds | Work
H 0 0 20.1386974 0.00000 100% - 0s
Explored 0 nodes (377 simplex iterations) in 0.08 seconds
Optimal solution found (tolerance 0.00e+00)
Solving fixed MIQCP.
Barrier solved model in 10 iterations and 0.01 seconds
Solving KKT system to obtain QCP duals...
Optimize a model with 80 rows, 126 columns and 240 nonzeros
Barrier solved model in 16 iterations and 0.02 seconds
Fixed MIQCP status(2): Model was solved to optimality (subject to tolerances).
MIQCP Solution: 20.138697 (377 iterations, 0 nodes)
Best possible: 20.138697
This looks all fine. Except the correct (global) optimal objective is 2.2388. The reason is: this MIQCP model is actually non-convex. Gurobi usually rejects those models (with a message about a Q matrix not being positive definite) but in this case it slips through. The solution is probably just a local optimum.