In your equation for \$R_{eq}\$ you have assumed that \$R_{19}\$, \$R_{21}\$, \$R_{22}\$ and \$R_{24}\$ are in series, which is incorrect:
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/0F2s1.png)
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/y9RcX.png)
These resistors form a network which cannot be treated using the usual formulae for series and parallel connected resistors.
Δ-Y (Delta-Wye) Transform
If you identify something like this "delta" or "pi" formation
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/2CzV9.png)
you may transform it into a functionally identical "wye" or "tee" form:
![schematic](https://cdn.statically.io/img/i.sstatic.net/vsapq.png)
simulate this circuit – Schematic created using CircuitLab
Everything you need to know about how to make this transformation is covered in the Wikipedia article on the subject: Wye-Delta Transform
After applying that transformation, you will have a circuit that looks like this:
![schematic](https://cdn.statically.io/img/i.sstatic.net/FlK08.png)
simulate this circuit
As you can see, \$R_z\$ and \$R_{24}\$ are in series, and can be replaced by single resistor of value \$R_z + R_{24}\$. It may be tempting to replace the series-connected pair \$R_y\$ and \$R_{23}\$ in the same manner, but you still require node A3 to be exposed, to find the Thevenin equivalent.
Y-Δ (Wye-Delta) Transform
Alternatively, take the "wye" formation of \$R_{21}\$, \$R_{22}\$ and \$R_{24}\$, and convert them to a "delta":
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/bBdOm.png)
Following that you would arrive here:
![schematic](https://cdn.statically.io/img/i.sstatic.net/BAzpn.png)
simulate this circuit
Now you see that \$R_{xz}\$ is in parallel with \$R_{20}\$, \$R_{xy}\$ is in parallel with \$R_{19}\$ and \$R_{yz}\$ is in parallel with \$R_{23}\$, and all these combinations may be replaced by single resistances.
You still might not be out of the woods, though. Whichever route you take, you may still have remaining "delta" or "wye" networks to take care of, but they can all be solved by applying another transform, in the same way.
Essentially you keep identifying problematic resistor configurations, applying a wye-delta or delta-wye transformation in each case, and further simplifying. At some point you will arrive at an equivalent circuit which can be reduced using only the formulae for parallel and series connected resistors, and you have your solution.
Succesive Thévenin Equivalencies
Another way to find the overall Thévenin equivalent circuit, is to take the apporoach Jonk suggested in his answer. That is, to successively apply Thévenin's theorem to small subsections of the circuit.
You must identify known potentials and simple resistor combinations that are easy to convert to their combined Thévenin equivalent. This is most easily achieved by redrawing the circuit in a manner which exposes such sub-units. Here's the same circuit rearranged in such a way:
![schematic](https://cdn.statically.io/img/i.sstatic.net/qxu3L.png)
simulate this circuit
\$R_{20}\$, being directly across the voltage source, will play no role at all in the calculations for that equivalent circuit, and can be disregarded.
From this starting point, it's clear that \$R_{21}\$ and \$R_{24}\$ may be considered a simple potential divider with 10V at the top and 0V at the bottom. Convert this section to its Thévenin equivalent:
![schematic](https://cdn.statically.io/img/i.sstatic.net/9i3q9.png)
simulate this circuit
Repeat the operation for the section consisting of \$R_{T1}\$, \$R_{22}\$ and \$R_{19}\$:
![schematic](https://cdn.statically.io/img/i.sstatic.net/rbeNI.png)
Keep doing this until you've reduced the circuit to a single resistor and a single voltage source.