diff --git a/.github/workflows/publish_docs.yml b/.github/workflows/publish_docs.yml
index bb63c7b6..78979b96 100644
--- a/.github/workflows/publish_docs.yml
+++ b/.github/workflows/publish_docs.yml
@@ -1,4 +1,4 @@
-name: Publish Sphinx Documentation
+name: Docs
 
 on:
   push:
diff --git a/navlie/filters.py b/navlie/filters.py
index b7f06efe..0bd3739d 100644
--- a/navlie/filters.py
+++ b/navlie/filters.py
@@ -141,9 +141,8 @@ def predict(
 
         details_dict = {}
         if u is not None:
-            A = self.process_model.jacobian(x_jac, u, dt)
             Q = self.process_model.covariance(x_jac, u, dt)
-            x_new.state = self.process_model.evaluate(x.state, u, dt)
+            x_new.state, A = self.process_model.evaluate_with_jacobian(x.state, u, dt)
             x_new.covariance = A @ x.covariance @ A.T + Q
             x_new.symmetrize()
             x_new.state.stamp = t_km1 + dt
@@ -217,13 +216,14 @@ def correct(
 
         if x_jac is None:
             x_jac = x.state
-        y_check = y.model.evaluate(x.state)
+
+        y_check, G = y.model.evaluate_with_jacobian(x.state)
 
         details_dict = {}
         if y_check is not None:
             P = x.covariance
             R = np.atleast_2d(y.model.covariance(x_jac))
-            G = np.atleast_2d(y.model.jacobian(x_jac))
+            G = np.atleast_2d(G)
             z = y.minus(y_check)
             S = G @ P @ G.T + R
 
@@ -344,8 +344,8 @@ def correct(
                 x_op_jac = x_op
 
             R = np.atleast_2d(y.model.covariance(x_op))
-            G = np.atleast_2d(y.model.jacobian(x_op_jac))
-            y_check = y.model.evaluate(x_op)
+            y_check, G = y.model.evaluate_with_jacobian(x_op)
+            G = np.atleast_2d(G)
             z = y.minus(y_check)
             e = x_op.minus(x.state).reshape((-1, 1))
             J = x_op.plus_jacobian(e)
@@ -417,7 +417,8 @@ def correct(
                 x_op_jac = x_op
 
             # Re-evaluate the jacobians at our latest operating point
-            G = np.atleast_2d(y.model.jacobian(x_op_jac))
+            _, G = y.model.evaluate_with_jacobian(x_op_jac)
+            G = np.atleast_2d(G)
             R = np.atleast_2d(y.model.covariance(x_op_jac))
             e = x_op.minus(x.state).reshape((-1, 1))
             J = x_op.plus_jacobian(e)
@@ -703,7 +704,6 @@ def correct(
         R = y.model.covariance(x.state)
 
         n_x = x.state.dof
-        n_y = y.value.size
 
         if (n_x, self.method) in self._sigmapoint_cache:
             unit_sigmapoints, w = self._sigmapoint_cache[(n_x, self.method)]
@@ -714,31 +714,22 @@ def correct(
         P_sqrt = np.linalg.cholesky(P_xx)
         sigmapoints = P_sqrt @ unit_sigmapoints
 
-        n_sig = w.size
-
         y_check = y.model.evaluate(x.state)
 
         if y_check is not None:
-            y_propagated = [
+            y_propagated = np.array([
                 y.model.evaluate(x.state.plus(sp)).ravel()
                 for sp in sigmapoints.T
-            ]
+            ])
 
             # predicted measurement mean
-            y_mean = np.zeros(n_y)
-            for i in range(n_sig):
-                y_mean += w[i] * y_propagated[i]
+            y_mean = np.sum(w[:,None]*y_propagated, axis=0)
 
             # compute covariance of innovation and cross covariance
-            Pyy = np.zeros((n_y, n_y))
-            Pxy = np.zeros((n_x, n_y))
-            for i in range(n_sig):
-                err = y_propagated[i].reshape((-1, 1)) - y_mean.reshape((-1, 1))
-
-                Pyy += w[i] * err @ err.T
-                Pxy += w[i] * sigmapoints[:, i].reshape((-1, 1)) @ err.T
-
-            Pyy += R
+            err = y_propagated - y_mean
+            sigmapoints = sigmapoints.T
+            Pyy = np.sum(w[:,None,None]*err[:,:,None]*err[:,None,:], axis=0) + R
+            Pxy = np.sum(w[:,None,None]*sigmapoints[:,:,None]*err[:,None,:], axis=0)
 
             z = y.minus(y_mean)
 
diff --git a/navlie/types.py b/navlie/types.py
index 9d09b061..e3dabe7d 100644
--- a/navlie/types.py
+++ b/navlie/types.py
@@ -245,7 +245,17 @@ def jacobian(self, x: State) -> np.ndarray:
         :math:`\mathbf{G} = \partial \mathbf{g}(\mathbf{x})/ \partial \mathbf{x}`.
         """
         return self.jacobian_fd(x)
-
+    
+    def evaluate_with_jacobian(self, x: State) -> (np.ndarray, np.ndarray):
+        """
+        Evaluates the measurement model and simultaneously returns the Jacobian
+        as its second output argument. This is useful to override for
+        performance reasons when the model evaluation and Jacobian have a lot of
+        common calculations, and it is more efficient to calculate them in the
+        same function call.
+        """
+        return self.evaluate(x), self.jacobian(x)
+    
     def jacobian_fd(self, x: State, step_size=1e-6):
         """
         Calculates the model jacobian with finite difference.
@@ -351,6 +361,16 @@ def jacobian(self, x: State, u: Input, dt: float) -> np.ndarray:
             Process model Jacobian with respect to the state :math:`\mathbf{F}`.
         """
         return self.jacobian_fd(x, u, dt)
+    
+    def evaluate_with_jacobian(self, x: State, u: Input, dt: float) -> (State, np.ndarray):
+        """
+        Evaluates the process model and simultaneously returns the Jacobian as 
+        its second output argument. This is useful to override for
+        performance reasons when the model evaluation and Jacobian have a lot of
+        common calculations, and it is more efficient to calculate them in the
+        same function call.
+        """
+        return self.evaluate(x, u, dt), self.jacobian(x, u, dt)
 
     def jacobian_fd(
         self, x: State, u: Input, dt: float, step_size=1e-6, *args, **kwargs
diff --git a/tests/test_examples.py b/tests/test_examples.py
index 642dcfda..4c6c1e79 100644
--- a/tests/test_examples.py
+++ b/tests/test_examples.py
@@ -42,7 +42,7 @@ def test_ex_inertial_nav():
     import ex_inertial_nav
 
 # def test_ex_imm_se3():
-#     import ex_imm_se3
+#     import ex_imm_se3 # TODO.
 
 def test_ex_imm_vector():
     from ex_imm_vector import main