Terminating optimization when leaving root node posts a different bestbound than the one in log.

Python 3 + Gurobi 9.5

I'm currently doing lower bound analysis in the root node. Since I'm only interested in the root I terminate the optimization using callbacks:

if where == GRB.Callback.MIP and m.cbGet(GRB.Callback.MIP_NODCNT) != 0:
    m.terminate()
    bst = m.cbGet(GRB.Callback.MIP_OBJBST)
    bnd = m.cbGet(GRB.Callback.MIP_OBJBND)
    gap = (bst - bnd) / bst * 100
    print(
        f"\nLeft Root node\n\Best objective {bst:.12e}, best bound {bnd:.12e}, gap {gap:.4f} %\n"
    )

But then, the log shows that there's a better lower bound which, in turn, would mean a lower gap than the one obtained when terminating the optimization (if this new lowerbound corresponds to one obtained in the root).

Root relaxation: objective 5.307926e+04, 1202 iterations, 0.02 seconds (0.05 work units)

    Nodes    |    Current Node    |     Objective Bounds      |     Work
 Expl Unexpl |  Obj  Depth IntInf | Incumbent    BestBd   Gap | It/Node Time

     0     0 53079.2556    0   92          - 53079.2556      -     -    0s
     0     0 66534.4216    0   94          - 66534.4216      -     -    0s
     0     0 67567.6001    0   68          - 67567.6001      -     -    0s
H    0     0                    77129.600000 67567.6001  12.4%     -    0s
H    0     0                    75067.200000 68443.9437  8.82%     -    0s
     0     0 68443.9437    0   81 75067.2000 68443.9437  8.82%     -    0s
H    0     0                    74847.200000 68443.9437  8.56%     -    0s
     0     0 68448.5698    0   82 74847.2000 68448.5698  8.55%     -    0s
     0     0 68587.7181    0   78 74847.2000 68587.7181  8.36%     -    0s
     0     0 68774.2004    0   87 74847.2000 68774.2004  8.11%     -    0s
     0     0 68777.0450    0   90 74847.2000 68777.0450  8.11%     -    0s
     0     0 68778.1575    0   91 74847.2000 68778.1575  8.11%     -    0s
     0     0 68999.7748    0   89 74847.2000 68999.7748  7.81%     -    0s
     0     0 69003.6911    0   88 74847.2000 69003.6911  7.81%     -    0s
     0     0 69004.7754    0   93 74847.2000 69004.7754  7.81%     -    1s
     0     0 69156.0564    0   93 74847.2000 69156.0564  7.60%     -    1s
     0     0 69175.7393    0   94 74847.2000 69175.7393  7.58%     -    1s
     0     0 69176.0974    0  109 74847.2000 69176.0974  7.58%     -    1s
     0     0 69187.0660    0   94 74847.2000 69187.0660  7.56%     -    1s
     0     0 69187.5403    0   94 74847.2000 69187.5403  7.56%     -    1s
     0     0 69201.6884    0  104 74847.2000 69201.6884  7.54%     -    1s
     0     0 69201.6884    0  104 74847.2000 69201.6884  7.54%     -    1s
     0     2 69201.6884    0  104 74847.2000 69201.6884  7.54%     -    1s

Left Root node
Best objective 7.484720000000e+04, best bound 6.920268589866e+04, gap 7.5414 %


Cutting planes:
  Implied bound: 12
  MIR: 29
  Flow cover: 35
  Zero half: 6
  RLT: 11
  Relax-and-lift: 1
  Lazy constraints: 110

Explored 3 nodes (5385 simplex iterations) in 1.74 seconds (1.42 work units)
Thread count was 16 (of 16 available processors)

Solution count 3: 74847.2 75067.2 77129.6 

Solve interrupted
Best objective 7.484720000000e+04, best bound 6.970075276348e+04, gap 6.8759%

Where does this new best bound come from? If I'm telling the solver to terminate when the MIP_NODCNT != 0 then it shouldn't solve any other node but the root, right?

Should I be using MIPSOL or MIPNODE instead of MIP? or another callback code? Since I realize that the amount of explored nodes is 3 instead of 1.

Thanks in advance!
Nicolas