diff --git a/.github/workflows/test_benchmark_collection_models.yml b/.github/workflows/test_benchmark_collection_models.yml
index 7b1568b1fc..0b14d73716 100644
--- a/.github/workflows/test_benchmark_collection_models.yml
+++ b/.github/workflows/test_benchmark_collection_models.yml
@@ -68,7 +68,7 @@ jobs:
     - name: Run Gradient Checks
       run: |
         pip install git+https://github.com/ICB-DCM/fiddy.git \
-          && cd tests/benchmark-models && pytest ./test_petab_benchmark.py
+          && cd tests/benchmark-models && pytest --durations=10 ./test_petab_benchmark.py
 
     # upload results
     - uses: actions/upload-artifact@v4
diff --git a/tests/benchmark-models/test_petab_benchmark.py b/tests/benchmark-models/test_petab_benchmark.py
index 82d27f85eb..7cf7480c4e 100644
--- a/tests/benchmark-models/test_petab_benchmark.py
+++ b/tests/benchmark-models/test_petab_benchmark.py
@@ -34,6 +34,7 @@
     "Crauste_CellSystems2017",
     "Fujita_SciSignal2010",
 }
+models = list(sorted(models))
 
 debug = False
 if debug:
@@ -62,6 +63,8 @@ def test_benchmark_gradient(model, scale):
         "Sneyd_PNAS2002",
         "Bertozzi_PNAS2020",
         "Okuonghae_ChaosSolitonsFractals2020",
+        "Giordano_Nature2020-False",
+        "Zheng_PNAS2012",
     ):
         # not really worth the effort trying to fix these cases if they
         # only fail on linear scale
@@ -145,7 +148,7 @@ def test_benchmark_gradient(model, scale):
         sizes=sizes,
         direction_ids=parameter_ids,
         method_ids=[MethodId.CENTRAL, MethodId.FORWARD, MethodId.BACKWARD],
-        success_checker=Consistency(atol=1e-4, rtol=0.2),
+        success_checker=Consistency(atol=1e-4, rtol=1e-1),
         expected_result=expected_derivative,
         relative_sizes=not scale,
     )