diff --git a/tests/benchmark-models/test_petab_benchmark.py b/tests/benchmark-models/test_petab_benchmark.py
index 7af50087d5..eb6bf1ccb1 100644
--- a/tests/benchmark-models/test_petab_benchmark.py
+++ b/tests/benchmark-models/test_petab_benchmark.py
@@ -101,8 +101,8 @@ class GradientCheckSettings:
     rng_seed=1,
 )
 settings["Sneyd_PNAS2002"] = GradientCheckSettings(
-    atol_sim=1e-16,
-    rtol_sim=1e-10,
+    atol_sim=1e-15,
+    rtol_sim=1e-12,
     atol_check=1e-5,
     rtol_check=1e-4,
 )
@@ -115,6 +115,11 @@ class GradientCheckSettings:
 settings["Brannmark_JBC2010"] = GradientCheckSettings(
     ss_sensitivity_mode=amici.SteadyStateSensitivityMode.integrationOnly,
 )
+settings["Borghans_BiophysChem1997"] = GradientCheckSettings(
+    rng_seed=2,
+    atol_check=1e-5,
+    rtol_check=1e-3,
+)
 
 
 def assert_gradient_check_success(
@@ -123,6 +128,7 @@ def assert_gradient_check_success(
     point: np.array,
     atol: float,
     rtol: float,
+    always_print: bool = False,
 ) -> None:
     if not derivative.df.success.all():
         raise AssertionError(
@@ -135,7 +141,7 @@ def assert_gradient_check_success(
     )
     check_result = check(rtol=rtol, atol=atol)
 
-    if check_result.success is True:
+    if check_result.success is True and not always_print:
         return
 
     df = check_result.df
@@ -144,12 +150,22 @@ def assert_gradient_check_success(
     max_adiff = df["abs_diff"].max()
     max_rdiff = df["rel_diff"].max()
     with pd.option_context("display.max_columns", None, "display.width", None):
+        message = (
+            f"Gradient check failed:\n{df}\n\n"
+            f"Maximum absolute difference: {max_adiff}\n"
+            f"Maximum relative difference: {max_rdiff}"
+        )
+
+    if check_result.success is False:
         raise AssertionError(
             f"Gradient check failed:\n{df}\n\n"
             f"Maximum absolute difference: {max_adiff}\n"
             f"Maximum relative difference: {max_rdiff}"
         )
 
+    if always_print:
+        print(message)
+
 
 @pytest.mark.filterwarnings(
     "ignore:divide by zero encountered in log",
@@ -205,7 +221,7 @@ def test_benchmark_gradient(model, scale, sensitivity_method, request):
     amici_solver = amici_model.getSolver()
     amici_solver.setAbsoluteTolerance(cur_settings.atol_sim)
     amici_solver.setRelativeTolerance(cur_settings.rtol_sim)
-    amici_solver.setMaxSteps(int(1e5))
+    amici_solver.setMaxSteps(2 * 10**5)
     amici_solver.setSensitivityMethod(sensitivity_method)
     # TODO: we should probably test all sensitivity modes
     amici_model.setSteadyStateSensitivityMode(cur_settings.ss_sensitivity_mode)
@@ -278,6 +294,7 @@ def test_benchmark_gradient(model, scale, sensitivity_method, request):
         point,
         rtol=cur_settings.rtol_check,
         atol=cur_settings.atol_check,
+        always_print=True,
     )