To solve the optimization problem described, we first need to convert the natural language description into a symbolic representation. Let's define:

- $x_1$ as the number of batches of pancakes.
- $x_2$ as the number of batches of pretzels.

The objective function is to maximize profit. Given that the profit per batch is $25 for pancakes and $50 for pretzels, we can represent the total profit as:

\[ \text{Maximize: } 25x_1 + 50x_2 \]

The constraints based on the available capacity are:

1. Oven time constraint: \(2.5x_1 + 1.5x_2 \leq 65\)
2. Pastry chef time constraint: \(0.5x_1 + 3x_2 \leq 35\)

Additionally, we have non-negativity constraints since the number of batches cannot be negative:

\[ x_1 \geq 0, x_2 \geq 0 \]

So, the symbolic representation in JSON format is:
```json
{
    'sym_variables': [('x1', 'number of batches of pancakes'), ('x2', 'number of batches of pretzels')],
    'objective_function': '25*x1 + 50*x2',
    'constraints': ['2.5*x1 + 1.5*x2 <= 65', '0.5*x1 + 3*x2 <= 35', 'x1 >= 0', 'x2 >= 0']
}
```

Now, let's implement this problem in Gurobi using Python:
```python
from gurobipy import *

# Create a new model
m = Model("Tom_Bakery")

# Define the variables
x1 = m.addVar(vtype=GRB.CONTINUOUS, name="pancakes", lb=0)
x2 = m.addVar(vtype=GRB.CONTINUOUS, name="pretzels", lb=0)

# Set the objective function
m.setObjective(25*x1 + 50*x2, GRB.MAXIMIZE)

# Add constraints
m.addConstr(2.5*x1 + 1.5*x2 <= 65, "oven_time")
m.addConstr(0.5*x1 + 3*x2 <= 35, "pastry_chef_time")

# Optimize the model
m.optimize()

# Print the results
if m.status == GRB.OPTIMAL:
    print("Optimal solution found:")
    print(f"Number of batches of pancakes: {x1.x}")
    print(f"Number of batches of pretzels: {x2.x}")
    print(f"Maximum profit: ${m.objVal:.2f}")
else:
    print("No optimal solution found")
```