diff --git a/conda/environments/all_cuda-129_arch-aarch64.yaml b/conda/environments/all_cuda-129_arch-aarch64.yaml
index 96adb871d2..140479f536 100644
--- a/conda/environments/all_cuda-129_arch-aarch64.yaml
+++ b/conda/environments/all_cuda-129_arch-aarch64.yaml
@@ -71,7 +71,6 @@ dependencies:
 - rmm==26.2.*,>=0.0.0a0
 - scikit-build-core>=0.10.0
 - scikit-learn>=1.4
-- scikit-learn>=1.4,<1.8.0
 - scipy>=1.11.0
 - seaborn
 - sphinx
diff --git a/conda/environments/all_cuda-129_arch-x86_64.yaml b/conda/environments/all_cuda-129_arch-x86_64.yaml
index 7839dc67b7..5f8e8903a6 100644
--- a/conda/environments/all_cuda-129_arch-x86_64.yaml
+++ b/conda/environments/all_cuda-129_arch-x86_64.yaml
@@ -71,7 +71,6 @@ dependencies:
 - rmm==26.2.*,>=0.0.0a0
 - scikit-build-core>=0.10.0
 - scikit-learn>=1.4
-- scikit-learn>=1.4,<1.8.0
 - scipy>=1.11.0
 - seaborn
 - sphinx
diff --git a/conda/environments/all_cuda-130_arch-aarch64.yaml b/conda/environments/all_cuda-130_arch-aarch64.yaml
index 494da2e884..449b5ec714 100644
--- a/conda/environments/all_cuda-130_arch-aarch64.yaml
+++ b/conda/environments/all_cuda-130_arch-aarch64.yaml
@@ -71,7 +71,6 @@ dependencies:
 - rmm==26.2.*,>=0.0.0a0
 - scikit-build-core>=0.10.0
 - scikit-learn>=1.4
-- scikit-learn>=1.4,<1.8.0
 - scipy>=1.11.0
 - seaborn
 - sphinx
diff --git a/conda/environments/all_cuda-130_arch-x86_64.yaml b/conda/environments/all_cuda-130_arch-x86_64.yaml
index 3e8908c335..6b2b6754d1 100644
--- a/conda/environments/all_cuda-130_arch-x86_64.yaml
+++ b/conda/environments/all_cuda-130_arch-x86_64.yaml
@@ -71,7 +71,6 @@ dependencies:
 - rmm==26.2.*,>=0.0.0a0
 - scikit-build-core>=0.10.0
 - scikit-learn>=1.4
-- scikit-learn>=1.4,<1.8.0
 - scipy>=1.11.0
 - seaborn
 - sphinx
diff --git a/dependencies.yaml b/dependencies.yaml
index f77919dd8d..dc71360a08 100644
--- a/dependencies.yaml
+++ b/dependencies.yaml
@@ -486,7 +486,7 @@ dependencies:
           - pytest-cov
           - pytest-xdist
           - seaborn
-          - scikit-learn>=1.4,<1.8.0
+          - *scikit_learn
           - statsmodels
           - tenacity
           - umap-learn==0.5.7
diff --git a/python/cuml/cuml_accel_tests/integration/test_hdbscan.py b/python/cuml/cuml_accel_tests/integration/test_hdbscan.py
index 68b80205a3..c7c01efd81 100644
--- a/python/cuml/cuml_accel_tests/integration/test_hdbscan.py
+++ b/python/cuml/cuml_accel_tests/integration/test_hdbscan.py
@@ -13,6 +13,9 @@
 from sklearn.preprocessing import StandardScaler
 
 if Version(sklearn.__version__) >= Version("1.8.0.dev0"):
+    # NOTE: Remove this skip when issue
+    # https://github.com/scikit-learn-contrib/hdbscan/issues/689 is resolved,
+    # as it blocks compatibility with scikit-learn >= 1.8.0.dev0.
     pytest.skip(
         "hdbscan requires sklearn < 1.8.0.dev0", allow_module_level=True
     )
diff --git a/python/cuml/cuml_accel_tests/upstream/scikit-learn/xfail-list.yaml b/python/cuml/cuml_accel_tests/upstream/scikit-learn/xfail-list.yaml
index 421305135c..1629121a09 100644
--- a/python/cuml/cuml_accel_tests/upstream/scikit-learn/xfail-list.yaml
+++ b/python/cuml/cuml_accel_tests/upstream/scikit-learn/xfail-list.yaml
@@ -1,3 +1,44 @@
+- reason: AUC standard deviation differs slightly with cuml.accel in sklearn 1.8
+  marker: cuml_accel_bugs
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-False-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-False-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-True-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-True-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-False-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-False-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-True-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-True-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-False-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-False-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-True-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-True-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-False-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-False-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-True-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-True-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[None-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[None-curve_kwargs1]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[single-None]"
+  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[single-curve_kwargs1]"
+- reason: Search CV sample weight equivalence differs with cuml.accel in sklearn 1.8
+  marker: cuml_accel_bugs
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.model_selection.tests.test_search::test_search_cv_sample_weight_equivalence[estimator0]"
+- reason: Test should fail with cuml.accel
+  marker: cuml_accel_bugs
+  condition: scikit-learn<1.8
+  tests:
+  - "sklearn.linear_model.tests.test_coordinate_descent::test_enet_coordinate_descent[Lasso-1-kwargs1]"
+  - "sklearn.linear_model.tests.test_coordinate_descent::test_warm_start_convergence"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-6-24-False-ElasticNet]"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-6-24-True-ElasticNet]"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-6-24-False-ElasticNet]"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-6-24-True-ElasticNet]"
+  - "sklearn.manifold.tests.test_t_sne::test_pca_initialization_not_compatible_with_sparse_input[csr_array]"
+  - "sklearn.manifold.tests.test_t_sne::test_pca_initialization_not_compatible_with_sparse_input[csr_matrix]"
 - reason: Test should fail with cuml.accel
   marker: cuml_accel_bugs
   tests:
@@ -200,7 +241,6 @@
   - "sklearn.linear_model.tests.test_common::test_balance_property[42-True-LogisticRegression]"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_check_input_false"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_elasticnet_precompute_gram_weighted_samples"
-  - "sklearn.linear_model.tests.test_coordinate_descent::test_enet_coordinate_descent[Lasso-1-kwargs1]"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_enet_copy_X_False_check_input_False"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_enet_float_precision"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_enet_multitarget"
@@ -215,7 +255,6 @@
   - "sklearn.linear_model.tests.test_coordinate_descent::test_lasso_readonly_data"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_lasso_toy"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_lasso_zero"
-  - "sklearn.linear_model.tests.test_coordinate_descent::test_warm_start_convergence"
   - "sklearn.linear_model.tests.test_coordinate_descent::test_warm_start_convergence_with_regularizer_decrement"
   - "sklearn.linear_model.tests.test_ransac::test_perfect_horizontal_line"
   - "sklearn.linear_model.tests.test_ransac::test_ransac_exceed_max_skips"
@@ -240,17 +279,13 @@
   - "sklearn.linear_model.tests.test_sag::test_step_size_alpha_error"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_same_multiple_output_sparse_dense[coo_array]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_same_multiple_output_sparse_dense[coo_matrix]"
-  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-6-24-False-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-6-24-False-Lasso]"
-  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-6-24-True-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-6-24-True-Lasso]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-True-6-24-False-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-True-6-24-False-Lasso]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-True-6-24-True-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-True-6-24-True-Lasso]"
-  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-6-24-False-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-6-24-False-Lasso]"
-  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-6-24-True-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-6-24-True-Lasso]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-True-6-24-False-ElasticNet]"
   - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-True-6-24-False-Lasso]"
@@ -262,8 +297,6 @@
   - "sklearn.manifold.tests.test_t_sne::test_bh_match_exact"
   - "sklearn.manifold.tests.test_t_sne::test_binary_perplexity_stability"
   - "sklearn.manifold.tests.test_t_sne::test_n_iter_without_progress"
-  - "sklearn.manifold.tests.test_t_sne::test_pca_initialization_not_compatible_with_sparse_input[csr_array]"
-  - "sklearn.manifold.tests.test_t_sne::test_pca_initialization_not_compatible_with_sparse_input[csr_matrix]"
   - "sklearn.manifold.tests.test_t_sne::test_preserve_trustworthiness_approximately[pca-barnes_hut]"
   - "sklearn.manifold.tests.test_t_sne::test_preserve_trustworthiness_approximately[pca-exact]"
   - "sklearn.manifold.tests.test_t_sne::test_preserve_trustworthiness_approximately[random-barnes_hut]"
@@ -451,11 +484,6 @@
   - "sklearn.neighbors.tests.test_neighbors::test_neighbors_metrics[float64-minkowski]"
   - "sklearn.preprocessing.tests.test_data::test_standard_scaler_partial_fit_numerical_stability[csc_array]"
   - "sklearn.preprocessing.tests.test_data::test_standard_scaler_partial_fit_numerical_stability[csr_array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-csr_array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-csr_matrix]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-log-csr_array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-log-csr_matrix]"
   - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[False-isotonic]"
   - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[False-sigmoid]"
   - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[True-isotonic]"
@@ -781,11 +809,6 @@
   - "sklearn.model_selection.tests.test_classification_threshold::test_fit_and_score_over_thresholds_sample_weight"
   - "sklearn.model_selection.tests.test_classification_threshold::test_tuned_threshold_classifier_cv_zeros_sample_weights_equivalence"
   - "sklearn.model_selection.tests.test_validation::test_cross_val_predict_class_subset"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-csr_array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-csr_matrix]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-log-csr_array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-log-csr_matrix]"
   - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[False-isotonic]"
   - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[False-sigmoid]"
   - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[True-isotonic]"
@@ -834,31 +857,21 @@
   - "sklearn.tests.test_common::test_estimators[LogisticRegression()-check_sample_weight_equivalence_on_sparse_data]"
   - "sklearn.tests.test_common::test_estimators[Ridge()-check_non_transformer_estimators_n_iter]"
   - "sklearn.utils.tests.test_estimator_checks::test_check_estimator_pairwise"
-- reason: Test should fail with cuml.accel (scikit-learn 1.7)
+- reason: Test should fail with cuml.accel (scikit-learn <1.8)
   marker: cuml_accel_bugs
-  condition: scikit-learn == 1.7.*
+  condition: scikit-learn<1.8
   tests:
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-False-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-False-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-True-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-decision_function-True-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-False-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-False-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-True-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[False-False-predict_proba-True-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-False-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-False-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-True-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-decision_function-True-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-False-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-False-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-True-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_display_plotting_from_cv_results[True-False-predict_proba-True-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[None-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[None-curve_kwargs1]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[single-None]"
-  - "sklearn.metrics._plot.tests.test_roc_curve_display::test_roc_curve_from_cv_results_legend_label[single-curve_kwargs1]"
-  - "sklearn.model_selection.tests.test_search::test_search_cv_sample_weight_equivalence[estimator0]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-array]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-csr_array]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-log-csr_matrix]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-log-csr_array]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-log-csr_matrix]"
+- reason: Test should fail with cuml.accel (scikit-learn<1.6)
+  marker: cuml_accel_bugs
+  condition: scikit-learn<1.6
+  tests:
+  - "sklearn.tests.test_common::test_check_n_features_in_after_fitting[LogisticRegression()]"
+  - "sklearn.tests.test_common::test_pandas_column_name_consistency[HalvingGridSearchCV(cv=2,estimator=LogisticRegression(),min_resources='smallest',param_grid={'C':[0.1,1.0]},random_state=0)0]"
 - reason: Test should fail with cuml.accel (scikit-learn<1.7)
   marker: cuml_accel_bugs
   condition: scikit-learn<1.5
@@ -869,12 +882,6 @@
   - "sklearn.decomposition.tests.test_pca::test_sparse_pca_solver_error[42-csr_array-auto]"
   - "sklearn.decomposition.tests.test_pca::test_sparse_pca_solver_error[42-csr_matrix-auto]"
   - "sklearn.manifold.tests.test_t_sne::test_n_iter_used"
-- reason: Test should fail with cuml.accel (scikit-learn<1.7)
-  marker: cuml_accel_bugs
-  condition: scikit-learn<1.7
-  tests:
-  - "sklearn.tests.test_common::test_check_n_features_in_after_fitting[LogisticRegression()]"
-  - "sklearn.tests.test_common::test_pandas_column_name_consistency[HalvingGridSearchCV(cv=2,estimator=LogisticRegression(),min_resources='smallest',param_grid={'C':[0.1,1.0]},random_state=0)0]"
 - reason: Test should fail with cuml.accel (scikit-learn<1.7)
   marker: cuml_accel_bugs
   condition: scikit-learn<1.7,>=1.5.0
@@ -906,6 +913,11 @@
   strict: false
   tests:
   - "sklearn.tests.test_common::test_check_n_features_in_after_fitting[SpectralEmbedding()]"
+- reason: cuml raises a different error if X doesn't have expected n features
+  marker: cuml_accel_check_n_features_in
+  condition: scikit-learn==1.7.*
+  tests:
+  - "sklearn.tests.test_common::test_estimators[SVC()-check_n_features_in_after_fitting]"
 - reason: cuml raises a different error if X doesn't have expected n features
   marker: cuml_accel_check_n_features_in
   condition: scikit-learn>=1.6
@@ -923,7 +935,6 @@
   - "sklearn.tests.test_common::test_estimators[RandomForestClassifier()-check_n_features_in_after_fitting]"
   - "sklearn.tests.test_common::test_estimators[RandomForestRegressor()-check_n_features_in_after_fitting]"
   - "sklearn.tests.test_common::test_estimators[Ridge()-check_n_features_in_after_fitting]"
-  - "sklearn.tests.test_common::test_estimators[SVC()-check_n_features_in_after_fitting]"
   - "sklearn.tests.test_common::test_estimators[SVR()-check_n_features_in_after_fitting]"
   - "sklearn.tests.test_common::test_estimators[TruncatedSVD()-check_n_features_in_after_fitting]"
 - reason: cuml doesn't set `feature_names_in_` properly
@@ -955,6 +966,15 @@
   tests:
   - "sklearn.mixture.tests.test_gaussian_mixture::test_gaussian_mixture_precisions_init_diag"
   - "sklearn.utils.tests.test_estimator_html_repr::test_show_arrow_pipeline"
+- reason: Test is flaky with cuml.accel
+  marker: cuml_accel_flaky
+  condition: scikit-learn<1.8
+  strict: false
+  tests:
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-barnes_hut]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-exact]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-barnes_hut]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-exact]"
 - reason: Test is flaky with cuml.accel
   marker: cuml_accel_flaky
   condition: scikit-learn>=1.6
@@ -983,10 +1003,6 @@
   - "sklearn.feature_selection.tests.test_sequential::test_unsupervised_model_fit[2]"
   - "sklearn.feature_selection.tests.test_sequential::test_unsupervised_model_fit[3]"
   - "sklearn.manifold.tests.test_spectral_embedding::test_pipeline_spectral_clustering"
-  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-barnes_hut]"
-  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-exact]"
-  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-barnes_hut]"
-  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-exact]"
   - "sklearn.manifold.tests.test_t_sne::test_optimization_minimizes_kl_divergence"
   - "sklearn.manifold.tests.test_t_sne::test_uniform_grid[barnes_hut]"
   - "sklearn.manifold.tests.test_t_sne::test_uniform_grid[exact]"
@@ -1159,6 +1175,11 @@
   - "sklearn.linear_model._glm.tests.test_glm::test_linalg_warning_with_newton_solver[42]"
   - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_path_convergence_fail"
   - "sklearn.linear_model.tests.test_logistic::test_newton_cholesky_fallback_to_lbfgs[42]"
+- reason: SVC input handling and validation
+  marker: cuml_accel_svc_estimator_checks
+  condition: scikit-learn<1.8
+  tests:
+  - "sklearn.svm.tests.test_svm::test_svc_nonfinite_params"
 - reason: SVC input handling and validation
   marker: cuml_accel_svc_estimator_checks
   condition: scikit-learn>=1.6
@@ -1168,7 +1189,6 @@
 - reason: SVC input handling and validation
   marker: cuml_accel_svc_estimator_checks
   tests:
-  - "sklearn.svm.tests.test_svm::test_svc_nonfinite_params"
   - "sklearn.tests.test_common::test_estimators[SVC()-check_classifier_data_not_an_array]"
   - "sklearn.tests.test_common::test_estimators[SVC()-check_complex_data]"
   - "sklearn.tests.test_common::test_estimators[SVC()-check_estimators_nan_inf]"
@@ -1349,10 +1369,47 @@
   - "sklearn.tests.test_common::test_estimators[TSNE()-check_fit2d_predict1d]"
   - "sklearn.tests.test_common::test_estimators[TSNE()-check_methods_sample_order_invariance]"
   - "sklearn.tests.test_common::test_estimators[TSNE()-check_methods_subset_invariance]"
+- reason: Calibration temperature scaling differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[False-temperature]"
+  - "sklearn.tests.test_calibration::test_calibrated_classifier_cv_double_sample_weights_equivalence[True-temperature]"
+- reason: Elasticnet scores attribute layout differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_scores_attribute_layout_elasticnet"
 - reason: Flaky deviations in n_iter_ values in cuml.accel
   strict: false
   tests:
   - "sklearn.cluster.tests.test_k_means::test_kmeans_elkan_results[42-1e-100-dense-blobs]"
+- reason: Linear SVM sample weight handling differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVC-params0]"
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVC-params1]"
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVC-params2]"
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVC-params3]"
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVR-params4]"
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVR-params5]"
+  - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[42-LinearSVR-params6]"
+- reason: LinearSVC parameter validation differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[42-False-l1-hinge]"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[42-False-l2-hinge]"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[42-True-l1-hinge]"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[42-True-l1-squared_hinge]"
+- reason: LinearSVM
+  condition: scikit-learn<1.8
+  tests:
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-squared_hinge-array]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-squared_hinge-array]"
+  - "sklearn.svm.tests.test_svm::test_liblinear_set_coef"
+  - "sklearn.svm.tests.test_svm::test_linear_svm_convergence_warnings"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[False-l1-hinge]"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[False-l2-hinge]"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[True-l1-hinge]"
+  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[True-l1-squared_hinge]"
 - reason: LinearSVM
   condition: scikit-learn<=1.6
   tests:
@@ -1377,20 +1434,12 @@
   - "sklearn.feature_selection.tests.test_rfe::test_rfe_wrapped_estimator[RFECV-4-importance_getter0]"
   - "sklearn.feature_selection.tests.test_rfe::test_rfe_wrapped_estimator[RFECV-4-regressor_.coef_]"
   - "sklearn.model_selection.tests.test_search::test_grid_search_no_score"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-multi-class-squared_hinge-array]"
-  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-two-classes-squared_hinge-array]"
   - "sklearn.svm.tests.test_sparse::test_linearsvc[lil_array-dok_array]"
   - "sklearn.svm.tests.test_sparse::test_linearsvc[lil_matrix-dok_matrix]"
   - "sklearn.svm.tests.test_sparse::test_linearsvc_iris[csr_array]"
   - "sklearn.svm.tests.test_sparse::test_linearsvc_iris[csr_matrix]"
   - "sklearn.svm.tests.test_sparse::test_sparse_liblinear_intercept_handling"
   - "sklearn.svm.tests.test_svm::test_dense_liblinear_intercept_handling"
-  - "sklearn.svm.tests.test_svm::test_liblinear_set_coef"
-  - "sklearn.svm.tests.test_svm::test_linear_svm_convergence_warnings"
-  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[False-l1-hinge]"
-  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[False-l2-hinge]"
-  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[True-l1-hinge]"
-  - "sklearn.svm.tests.test_svm::test_linearsvc_parameters[True-l1-squared_hinge]"
   - "sklearn.tests.test_calibration::test_calibration_default_estimator"
   - "sklearn.tests.test_calibration::test_calibration_inconsistent_prefit_n_features_in"
   - "sklearn.tests.test_calibration::test_calibration_multiclass[1-False-isotonic]"
@@ -1420,10 +1469,12 @@
   - "sklearn.tests.test_common::test_pandas_column_name_consistency[LinearSVC()]"
   - "sklearn.tests.test_common::test_pandas_column_name_consistency[LinearSVR()]"
 - reason: LinearSVM test expects exact results on small data, which isn't guaranteed
+  condition: scikit-learn<1.8
   tests:
   - "sklearn.svm.tests.test_svm::test_bad_input[lil_array]"
   - "sklearn.svm.tests.test_svm::test_bad_input[lil_matrix]"
 - reason: LinearSVM tests too strict of tolerances due to differences in solvers
+  condition: scikit-learn<1.8
   tests:
   - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[LinearSVC-params0]"
   - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[LinearSVC-params1]"
@@ -1432,6 +1483,58 @@
   - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[LinearSVR-params4]"
   - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[LinearSVR-params5]"
   - "sklearn.svm.tests.test_svm::test_linearsvm_liblinear_sample_weight[LinearSVR-params6]"
+- reason: LogisticRegression liblinear sample weight handling differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_logisticregression_liblinear_sample_weight[42-params0]"
+  - "sklearn.linear_model.tests.test_logistic::test_logisticregression_liblinear_sample_weight[42-params1]"
+  - "sklearn.linear_model.tests.test_logistic::test_logisticregression_liblinear_sample_weight[42-params2]"
+- reason: LogisticRegression missing _predict_proba_lr attribute with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_logreg_predict_proba_multinomial[42]"
+- reason: LogisticRegression multiclass solvers differ with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_solvers_multiclass[True]"
+- reason: LogisticRegression sample weights handling differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_sample_weights[42-lbfgs-single]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_sample_weights[42-liblinear-single]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_sample_weights[42-newton-cg-single]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_sample_weights[42-newton-cholesky-single]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_sample_weights[42-sag-single]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_regression_sample_weights[42-saga-single]"
+- reason: LogisticRegressionCV fold coefficients differ with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_LogisticRegressionCV_on_folds"
+- reason: Multinomial logistic regression class weight handling differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_multinomial_logistic_regression_string_inputs"
+- reason: 'Numerical precision difference: cuML uses float32, test expects float64 precision'
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_cv[42-False]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_cv[42-True]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_cv_multinomial_score[42-neg_log_loss-multiclass_agg_list3]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_glmnet[lbfgs]"
+  - "sklearn.linear_model.tests.test_logistic::test_logistic_glmnet[newton-cholesky]"
+- reason: Numerical tolerance issue with Lasso sparse/dense equality in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-24-6-False-Lasso]"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_array-False-24-6-True-Lasso]"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-24-6-False-Lasso]"
+  - "sklearn.linear_model.tests.test_sparse_coordinate_descent::test_sparse_dense_equality[csc_matrix-False-24-6-True-Lasso]"
+- reason: Numerical tolerance issue with t-SNE sparse input in sklearn 1.8 (flaky)
+  condition: scikit-learn>=1.8
+  strict: false
+  tests:
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-random-exact]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-random-exact]"
 - reason: Ridge doesn't implement n_iter yet
   tests:
   - "sklearn.linear_model.tests.test_ridge::test_n_iter"
@@ -1473,3 +1576,122 @@
 - reason: The sklearn test has the error message accidentally flipped, our message is correct
   tests:
   - "sklearn.linear_model.tests.test_ridge::test_ridge_individual_penalties"
+- reason: cuML TSNE barnes_hut produces poor quality embeddings with sparse input
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-random-barnes_hut]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-random-barnes_hut]"
+- reason: cuML TSNE does not support sparse input with PCA initialization
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-pca-barnes_hut]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_array-pca-exact]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-pca-barnes_hut]"
+  - "sklearn.manifold.tests.test_t_sne::test_fit_transform_csr_matrix[csr_matrix-pca-exact]"
+- reason: cuML does not emit ConvergenceWarning in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-lbfgs-lbfgs failed to converge-max_iter0]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-lbfgs-lbfgs failed to converge-max_iter1]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-lbfgs-lbfgs failed to converge-max_iter2]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-lbfgs-lbfgs failed to converge-max_iter3]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-liblinear-Liblinear failed to converge, increase the number of iterations.-max_iter0]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-liblinear-Liblinear failed to converge, increase the number of iterations.-max_iter1]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-liblinear-Liblinear failed to converge, increase the number of iterations.-max_iter2]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-liblinear-Liblinear failed to converge, increase the number of iterations.-max_iter3]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-newton-cg-newton-cg failed to converge.* Increase the number of iterations.-max_iter0]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-newton-cg-newton-cg failed to converge.* Increase the number of iterations.-max_iter1]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-newton-cg-newton-cg failed to converge.* Increase the number of iterations.-max_iter2]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-newton-cg-newton-cg failed to converge.* Increase the number of iterations.-max_iter3]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-newton-cholesky-Newton solver did not converge after [0-9]* iterations-max_iter0]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-sag-The max_iter was reached which means the coef_ did not converge-max_iter0]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-sag-The max_iter was reached which means the coef_ did not converge-max_iter1]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-sag-The max_iter was reached which means the coef_ did not converge-max_iter2]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-sag-The max_iter was reached which means the coef_ did not converge-max_iter3]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-saga-The max_iter was reached which means the coef_ did not converge-max_iter0]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-saga-The max_iter was reached which means the coef_ did not converge-max_iter1]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-saga-The max_iter was reached which means the coef_ did not converge-max_iter2]"
+  - "sklearn.linear_model.tests.test_logistic::test_max_iter[42-saga-The max_iter was reached which means the coef_ did not converge-max_iter3]"
+  - "sklearn.svm.tests.test_svm::test_linear_svm_convergence_warnings[42]"
+- reason: cuML doesn't enforce sparse matrix int64 indices restriction
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_large_sparse_matrix[42-csr_array-liblinear]"
+  - "sklearn.linear_model.tests.test_logistic::test_large_sparse_matrix[42-csr_array-sag]"
+  - "sklearn.linear_model.tests.test_logistic::test_large_sparse_matrix[42-csr_array-saga]"
+  - "sklearn.linear_model.tests.test_logistic::test_large_sparse_matrix[42-csr_matrix-liblinear]"
+  - "sklearn.linear_model.tests.test_logistic::test_large_sparse_matrix[42-csr_matrix-sag]"
+  - "sklearn.linear_model.tests.test_logistic::test_large_sparse_matrix[42-csr_matrix-saga]"
+- reason: cuML doesn't support warm_start with newton solvers
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-1-False-newton-cg]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-1-False-newton-cholesky]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-1-True-newton-cg]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-1-True-newton-cholesky]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-inf-False-newton-cg]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-inf-False-newton-cholesky]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-inf-True-newton-cg]"
+  - "sklearn.linear_model.tests.test_logistic::test_warm_start_newton_solver[42-inf-True-newton-cholesky]"
+- reason: cuML doesn't validate NaN input in the same way as sklearn
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_nan"
+- reason: cuML proxy doesn't replicate sklearn warnings in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_l1_ratio_None_deprecated"
+  - "sklearn.linear_model.tests.test_logistic::test_l1_ratio_non_elasticnet"
+  - "sklearn.linear_model.tests.test_logistic::test_logisticregression_warns_with_n_jobs"
+  - "sklearn.linear_model.tests.test_logistic::test_lr_penalty_l1ratio_incompatible[l1-0.0]"
+  - "sklearn.linear_model.tests.test_logistic::test_lr_penalty_l1ratio_incompatible[l2-1.0]"
+  - "sklearn.linear_model.tests.test_logistic::test_penalty_deprecated[LogisticRegression]"
+  - "sklearn.linear_model.tests.test_logistic::test_penalty_none[42-lbfgs]"
+  - "sklearn.linear_model.tests.test_logistic::test_penalty_none[42-newton-cg]"
+  - "sklearn.linear_model.tests.test_logistic::test_penalty_none[42-newton-cholesky]"
+  - "sklearn.linear_model.tests.test_logistic::test_penalty_none[42-sag]"
+  - "sklearn.linear_model.tests.test_logistic::test_penalty_none[42-saga]"
+- reason: cuML proxy doesn't support direct writes to coef_/intercept_ attributes
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_write_parameters"
+- reason: cuML uses different solver backend, doesn't enforce liblinear restrictions
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_check_solver_option[LogisticRegression]"
+  - "sklearn.linear_model.tests.test_logistic::test_liblinear_multiclass_raises[LogisticRegression]"
+- reason: cuML uses float32 for liblinear solver, sklearn expects float64
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_dtype_match[csr_array-False-liblinear]"
+  - "sklearn.linear_model.tests.test_logistic::test_dtype_match[csr_array-True-liblinear]"
+  - "sklearn.linear_model.tests.test_logistic::test_dtype_match[csr_matrix-False-liblinear]"
+  - "sklearn.linear_model.tests.test_logistic::test_dtype_match[csr_matrix-True-liblinear]"
+- reason: cuPy sparse matrices don't support int64 dtype in decision_function output
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_predict_2_classes[csr_array]"
+  - "sklearn.linear_model.tests.test_logistic::test_predict_2_classes[csr_matrix]"
+  - "sklearn.linear_model.tests.test_logistic::test_predict_3_classes[csr_array]"
+  - "sklearn.linear_model.tests.test_logistic::test_predict_3_classes[csr_matrix]"
+- reason: cuml.accel handles sparse input differently in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.svm.tests.test_svm::test_bad_input[42-lil_array]"
+  - "sklearn.svm.tests.test_svm::test_bad_input[42-lil_matrix]"
+- reason: cuml.accel raises RuntimeError instead of ValueError for non-finite params in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.svm.tests.test_svm::test_svc_nonfinite_params[42]"
+- reason: l1_min_c calculation differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-log-csr_array]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-log-csr_matrix]"
+  - "sklearn.svm.tests.test_bounds::test_l1_min_c[no-intercept-squared_hinge-array]"
+- reason: liblinear solver behavior differs with cuml.accel in sklearn 1.8
+  condition: scikit-learn>=1.8
+  tests:
+  - "sklearn.linear_model.tests.test_logistic::test_liblinear_dual_random_state[42]"
+  - "sklearn.linear_model.tests.test_logistic::test_liblinear_with_large_values"
+  - "sklearn.svm.tests.test_svm::test_liblinear_set_coef[42]"
diff --git a/python/cuml/cuml_accel_tests/upstream/umap/run-tests.sh b/python/cuml/cuml_accel_tests/upstream/umap/run-tests.sh
index f2bc065042..3e0249b6bf 100755
--- a/python/cuml/cuml_accel_tests/upstream/umap/run-tests.sh
+++ b/python/cuml/cuml_accel_tests/upstream/umap/run-tests.sh
@@ -14,6 +14,21 @@ set -eu
 
 UMAP_TAG="release-0.5.7"
 
+# Skip tests for scikit-learn >= 1.8 -- umap-learn is not compatible with scikit-learn 1.8 yet
+python -c "
+import sys
+from packaging.version import Version
+import sklearn
+sys.exit(
+    int(
+        Version(sklearn.__version__) >= Version('1.8')
+    )
+)
+" || {
+    echo "Skipping umap tests for scikit-learn >= 1.8"
+    exit 0
+}
+
 THIS_DIRECTORY=$( cd "$(dirname "${BASH_SOURCE[0]}")" ; pwd -P )
 UMAP_REPO="${THIS_DIRECTORY}/umap-upstream"
 
diff --git a/python/cuml/pyproject.toml b/python/cuml/pyproject.toml
index 69e930c553..77569d1b6a 100644
--- a/python/cuml/pyproject.toml
+++ b/python/cuml/pyproject.toml
@@ -120,7 +120,7 @@ test = [
     "pytest-xdist",
     "pytest<9.0",
     "pyyaml",
-    "scikit-learn>=1.4,<1.8.0",
+    "scikit-learn>=1.4",
     "seaborn",
     "statsmodels",
     "tenacity",
diff --git a/python/cuml/tests/test_fil.py b/python/cuml/tests/test_fil.py
index 31997b52df..63c991702b 100644
--- a/python/cuml/tests/test_fil.py
+++ b/python/cuml/tests/test_fil.py
@@ -10,7 +10,9 @@
 import numpy as np
 import pandas as pd
 import pytest
+import sklearn
 import treelite
+from packaging.version import Version
 
 # Import XGBoost before scikit-learn to work around a libgomp bug
 # See https://github.com/dmlc/xgboost/issues/7110
@@ -910,6 +912,9 @@ def test_device_selection(device_id, model_kind, tmp_path):
         )
         xgb_model.fit(X, y)
         model_path = os.path.join(tmp_path, "xgb_class.ubj")
+        # skip with sklearn version 1.8.0.dev0
+        if Version(sklearn.__version__) >= Version("1.8.0.dev0"):
+            pytest.skip("xgboost is incompatible with sklearn >= 1.8.0.dev0")
         xgb_model.save_model(model_path)
         fm = ForestInference.load(
             model_path,
diff --git a/python/cuml/tests/test_hdbscan.py b/python/cuml/tests/test_hdbscan.py
index c6d876c8ac..ac3bfb8821 100644
--- a/python/cuml/tests/test_hdbscan.py
+++ b/python/cuml/tests/test_hdbscan.py
@@ -6,6 +6,8 @@
 import numpy as np
 import pandas as pd
 import pytest
+import sklearn
+from packaging.version import Version
 from pylibraft.common import DeviceResourcesSNMG
 from sklearn import datasets
 from sklearn.datasets import make_blobs
@@ -23,6 +25,11 @@
 from cuml.testing.datasets import make_pattern
 from cuml.testing.utils import array_equal
 
+if Version(sklearn.__version__) >= Version("1.8.0.dev0"):
+    pytest.skip(
+        "hdbscan requires sklearn < 1.8.0.dev0", allow_module_level=True
+    )
+
 dataset_names = ["noisy_circles", "noisy_moons", "varied"]
 
 
diff --git a/python/cuml/tests/test_linear_model.py b/python/cuml/tests/test_linear_model.py
index e4e243dedf..062d546675 100644
--- a/python/cuml/tests/test_linear_model.py
+++ b/python/cuml/tests/test_linear_model.py
@@ -329,7 +329,6 @@ def test_logistic_regression(
                 solver="saga",
                 C=C,
                 fit_intercept=fit_intercept,
-                multi_class="auto",
             )
         else:
             sklog = skLog(
@@ -337,7 +336,6 @@ def test_logistic_regression(
                 solver="saga",
                 C=C,
                 fit_intercept=fit_intercept,
-                multi_class="auto",
             )
     else:
         sklog = skLog(
@@ -345,7 +343,6 @@ def test_logistic_regression(
             solver="lbfgs",
             C=C,
             fit_intercept=fit_intercept,
-            multi_class="auto",
         )
 
     sklog.fit(X_train, y_train)
@@ -417,7 +414,7 @@ def test_logistic_regression_model_default(dtype):
     y_test = y_test.astype(dtype)
     culog = cuLog()
     culog.fit(X_train, y_train)
-    sklog = skLog(multi_class="auto")
+    sklog = skLog()
 
     sklog.fit(X_train, y_train)
 
@@ -586,11 +583,7 @@ def test_logistic_regression_predict_proba(
 
     sklog = skLog(
         fit_intercept=fit_intercept,
-        **(
-            {"solver": "lbfgs", "multi_class": "multinomial"}
-            if num_classes > 2
-            else {}
-        ),
+        **({"solver": "lbfgs"} if num_classes > 2 else {}),
     )
     sklog.coef_ = culog.coef_
     sklog.intercept_ = culog.intercept_ if fit_intercept else 0
diff --git a/python/cuml/tests/test_trustworthiness.py b/python/cuml/tests/test_trustworthiness.py
index e684f26442..668a9375df 100644
--- a/python/cuml/tests/test_trustworthiness.py
+++ b/python/cuml/tests/test_trustworthiness.py
@@ -4,6 +4,8 @@
 import cudf
 import numpy as np
 import pytest
+import sklearn
+from packaging.version import Version
 from sklearn.datasets import make_blobs
 from sklearn.manifold import trustworthiness as sklearn_trustworthiness
 from umap import UMAP
@@ -19,6 +21,10 @@
 @pytest.mark.filterwarnings(
     "ignore:n_jobs value.*overridden.*by setting random_state.*:UserWarning"
 )
+@pytest.mark.xfail(
+    condition=Version(sklearn.__version__) >= Version("1.8.0.dev0"),
+    reason="umap-learn is incompatible with sklearn >= 1.8.0",
+)
 def test_trustworthiness(
     input_type, n_samples, n_features, n_components, batch_size
 ):
diff --git a/python/cuml/tests/test_umap.py b/python/cuml/tests/test_umap.py
index a66c48dc34..4e80387de6 100644
--- a/python/cuml/tests/test_umap.py
+++ b/python/cuml/tests/test_umap.py
@@ -9,7 +9,9 @@
 import numpy as np
 import pytest
 import scipy.sparse as scipy_sparse
+import sklearn
 import umap
+from packaging.version import Version
 from pylibraft.common import DeviceResourcesSNMG
 from sklearn import datasets
 from sklearn.cluster import KMeans
@@ -29,6 +31,9 @@
     unit_param,
 )
 
+if Version(sklearn.__version__) >= Version("1.8.0.dev0"):
+    pytest.skip("umap requires sklearn < 1.8.0.dev0", allow_module_level=True)
+
 dataset_names = ["iris", "digits", "wine", "blobs"]