diff --git a/Jenkinsfile b/Jenkinsfile
index ff5fd03945..dc15599e0e 100644
--- a/Jenkinsfile
+++ b/Jenkinsfile
@@ -8,9 +8,6 @@ properties([
         booleanParam(defaultValue: true,
                      description: 'Whether to propagate commit status to GitHub',
                      name: 'propagateStatus'),
-        booleanParam(defaultValue: false,
-               description: 'If true, forces running pre-commit scope',
-               name: 'forceRunPrecommitScope'),
         string(defaultValue: '',
                description: 'Pipeline shared library version (branch/tag/commit). Determined automatically if empty',
                name: 'library_version')
diff --git a/pyproject.toml b/pyproject.toml
index d203f17947..d962982d24 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -59,8 +59,3 @@ requires = [
     "cmake~=3.24.0; platform_system == 'Darwin' and platform_machine == 'arm64'",
 ]
 build-backend = "py_build_cmake.build"
-
-[tool.pytest.ini_options]
-markers = [
-    "precommit: (deselect with '-m \"precommit\"')",
-]
diff --git a/src/cpp/src/lora/adapter.cpp b/src/cpp/src/lora/adapter.cpp
index 5335f15378..75655d4cde 100644
--- a/src/cpp/src/lora/adapter.cpp
+++ b/src/cpp/src/lora/adapter.cpp
@@ -100,10 +100,9 @@ struct AutoSafetensor: public safetensors_File {
 ConstantMap safetensor_to_constant_map(const ov::Tensor& safetensor) {
     AutoSafetensor safe_tensors_file{};
 
-    // Intentionally discard constness as safetensors_file_init requires a non-const pointer (used as read-only)
-    auto data_ptr = const_cast<char*>(safetensor.data<char>());
-    OPENVINO_ASSERT(safetensors_file_init(data_ptr, safetensor.get_byte_size(), &safe_tensors_file) == nullptr,
-                    "Cannot parse safetensor as a Safetensors file format. Safetensors file format is supported only");
+    OPENVINO_ASSERT(safetensors_file_init(safetensor.data<char>(), safetensor.get_byte_size(), &safe_tensors_file) == nullptr,
+        "Cannot parse safetensor as a Safetensors file format. Safetensors file format is supported only"
+    );
 
     ConstantMap tensors;
     for (int i = 0; i < safe_tensors_file.num_tensors; i++) {
diff --git a/src/cpp/src/speculative_decoding/speculative_decoding_stateful.cpp b/src/cpp/src/speculative_decoding/speculative_decoding_stateful.cpp
index 51dea61bec..9ee58ec62d 100644
--- a/src/cpp/src/speculative_decoding/speculative_decoding_stateful.cpp
+++ b/src/cpp/src/speculative_decoding/speculative_decoding_stateful.cpp
@@ -338,7 +338,7 @@ std::variant<int64_t, std::vector<int64_t>>
 
     auto sample_token = [&](const ov::Tensor& logits, std::size_t idx) {
         size_t sequence_offset = idx * vocab_size;
-        const float* logits_data = logits.data<float>() + sequence_offset;
+        float* logits_data = logits.data<float>() + sequence_offset;
         return std::max_element(logits_data, logits_data + vocab_size) - logits_data;
     };
 
diff --git a/tests/python_tests/README.md b/tests/python_tests/README.md
index b15737c2d2..de7b236652 100644
--- a/tests/python_tests/README.md
+++ b/tests/python_tests/README.md
@@ -14,29 +14,29 @@ pip install -r tests/python_tests/requirements.txt
 ## Run Tests
 
 ```sh
-python -m pytest tests/python_tests/ -m precommit
+python -m pytest tests/python_tests/
 ```
 
 If you have built GenAI library by yourself instead of using wheel please set `PYTHONPATH` so that test could find library, e.g.
 ```sh
-PYTHONPATH=$PYTHONPATH:.../openvino.genai/build-Release/ python -m pytest tests/python_tests/ -m precommit
+PYTHONPATH=$PYTHONPATH:.../openvino.genai/build-Release/ python -m pytest tests/python_tests/
 ```
 
 ## Customize tests run
 
-Tests have `precommit` set of models. `precommit` contains lightweight models which can be quickly inferred. If you wish to run specific tests, you can use `-k` option, for example to run only multibatch and chat tests:
+Tests have different sets of models for different purposes. If you wish to run specific tests, you can use `-k` option, for example to run only multibatch and chat tests:
 ```sh
-python -m pytest tests/python_tests/ -m precommit -k "test_multibatch and test_chat"
+python -m pytest tests/python_tests/ -k "test_multibatch and test_chat"
 ```
 
 If you wish to run all tests except beam search do the following:
 ```sh
-python -m pytest tests/python_tests/ -m precommit -k "not test_beam_search"
+python -m pytest tests/python_tests/ -k "not test_beam_search"
 ```
 
 Argument `--model_ids` can be used to run tests selectively only for specific models. HF model ids should be separated by space, e.g:
 ```sh
-python -m pytest tests/python_tests/ -m precommit -k "test_multibatch" --model_ids "TinyLlama/TinyLlama-1.1B-Chat-v1.0 Qwen/Qwen2-0.5B-Instruct"
+python -m pytest tests/python_tests/ -k "test_multibatch" --model_ids "TinyLlama/TinyLlama-1.1B-Chat-v1.0 Qwen/Qwen2-0.5B-Instruct"
 ```
 
 List of currently supported models can be found in tests/python_tests/models.py:get_models_list
diff --git a/tests/python_tests/conftest.py b/tests/python_tests/conftest.py
index 4044b61209..680715b6f2 100644
--- a/tests/python_tests/conftest.py
+++ b/tests/python_tests/conftest.py
@@ -53,6 +53,4 @@ def pytest_addoption(parser):
 
 
 def pytest_configure(config: pytest.Config):
-    marker = "precommit" if config.getoption("-m") == "precommit" else None
-    pytest.run_marker = marker
     pytest.selected_model_ids = config.getoption("--model_ids", default=None)
diff --git a/tests/python_tests/models/precommit b/tests/python_tests/models/lightweight
similarity index 100%
rename from tests/python_tests/models/precommit
rename to tests/python_tests/models/lightweight
diff --git a/tests/python_tests/pytest.ini b/tests/python_tests/pytest.ini
index 60b5cc39d8..f0807854c7 100644
--- a/tests/python_tests/pytest.ini
+++ b/tests/python_tests/pytest.ini
@@ -2,8 +2,8 @@
 
 markers =
     ; The following markers are defined for categorizing tests:
-    ; precommit   - Tests that should be run before committing code.
     ; real_models - Tests that involve execution of the models from models/real_models file
+    ; nightly     - Tests that should only run in nightly builds (uses large models or long running)
     ; samples     - Tests related to the sample models.
     ; llm         - Tests related to large language models.
     ; whisper     - Tests related to the Whisper model.
@@ -12,8 +12,8 @@ markers =
     ; vlm         - Tests related to the VLM model.
     ; rag         - Tests related to the RAG components.
     ; speech_generation - Tests related to text-to-speech generation
-    precommit
     real_models
+    nightly
     samples
     llm
     whisper
@@ -24,4 +24,4 @@ markers =
     rag
     speech_generation
 
-addopts = -m precommit
+addopts = -m "not real_models and not nightly"
diff --git a/tests/python_tests/samples/test_text2speech.py b/tests/python_tests/samples/test_text2speech.py
index 8b05c6530c..c497e3d65d 100644
--- a/tests/python_tests/samples/test_text2speech.py
+++ b/tests/python_tests/samples/test_text2speech.py
@@ -32,7 +32,6 @@ def teardown_class(self):
 
     @pytest.mark.speech_generation
     @pytest.mark.samples
-    @pytest.mark.precommit
     @pytest.mark.parametrize("convert_model", ["tiny-random-SpeechT5ForTextToSpeech"], indirect=True)
     @pytest.mark.parametrize("input_prompt", ["Hello everyone"])
     def test_sample_text_to_speech(self, convert_model, input_prompt):
@@ -54,7 +53,6 @@ def test_sample_text_to_speech(self, convert_model, input_prompt):
 
     @pytest.mark.speech_generation
     @pytest.mark.samples
-    @pytest.mark.precommit
     @pytest.mark.parametrize("convert_model", ["tiny-random-SpeechT5ForTextToSpeech"], indirect=True)
     @pytest.mark.parametrize("input_prompt", ["Test text to speech without speaker embedding file"])
     def test_sample_text_to_speech_no_speaker_embedding_file(self, convert_model, input_prompt):
diff --git a/tests/python_tests/test_continuous_batching.py b/tests/python_tests/test_continuous_batching.py
index 90ea7ac316..b32ba5e8e7 100644
--- a/tests/python_tests/test_continuous_batching.py
+++ b/tests/python_tests/test_continuous_batching.py
@@ -36,9 +36,8 @@ def read_models_list(file_name: str):
             models.append(model_name)
     return models
 
-@pytest.mark.precommit
-@pytest.mark.parametrize("model_id", read_models_list(os.path.join(os.path.dirname(os.path.realpath(__file__)), "models", "precommit")))
-def test_e2e_precommit(model_id):
+@pytest.mark.parametrize("model_id", read_models_list(os.path.join(os.path.dirname(os.path.realpath(__file__)), "models", "lightweight")))
+def test_e2e_lightweight_models(model_id):
     prompts, generation_configs = get_test_dataset()
     generate_and_compare(prompts=prompts,
                          generation_config=generation_configs,
@@ -73,7 +72,6 @@ def test_e2e_real_models(model_id):
 ]
 @pytest.mark.parametrize("generation_config", test_configs)
 @pytest.mark.parametrize("prompt", batched_prompts[1:])  # num_beams=15 diverges on the first prompt.
-@pytest.mark.precommit
 @pytest.mark.skip(reason="CVS-162891: Fix test_continuous_batching_vs_stateful tests after we started to compare cb vs sdpa")
 def test_continuous_batching_vs_stateful(prompt, generation_config):
     model_id = "facebook/opt-125m"
@@ -93,7 +91,6 @@ def test_continuous_batching_vs_stateful(prompt, generation_config):
 
 prompts = ['The Sun is yellow because', 'Difference between Jupiter and Mars is that', 'table is made of']
 @pytest.mark.parametrize("prompt", prompts)
-@pytest.mark.precommit
 def test_cb_streamer_vs_return_vs_stateful(prompt):
     model_id = "facebook/opt-125m"
     _, _, models_path = download_and_convert_model(model_id)
@@ -124,7 +121,6 @@ def test_cb_streamer_vs_return_vs_stateful(prompt):
 @pytest.mark.parametrize("input_type", [
     GenerationChatInputsType.STRING,
     GenerationChatInputsType.CHAT_HISTORY])
-@pytest.mark.precommit
 def test_chat_scenario_vs_stateful(model_id, generation_config_kwargs: dict, pipeline_type, input_type: GenerationChatInputsType):
     _, _, models_path = download_and_convert_model(model_id)
 
@@ -175,7 +171,6 @@ def test_chat_scenario_vs_stateful(model_id, generation_config_kwargs: dict, pip
 @pytest.mark.parametrize("generation_config_kwargs", generation_configs)
 @pytest.mark.parametrize("model_id", get_chat_models_list())
 @pytest.mark.parametrize("pipeline_type", [PipelineType.CONTINUOUS_BATCHING, PipelineType.SPECULATIVE_DECODING, PipelineType.PROMPT_LOOKUP_DECODING,])
-@pytest.mark.precommit
 def test_continuous_batching_add_request_health_check(model_id, generation_config_kwargs: dict, pipeline_type):
     _, _, models_path = download_and_convert_model(model_id)
 
@@ -206,7 +201,6 @@ def test_continuous_batching_add_request_health_check(model_id, generation_confi
 @pytest.mark.parametrize("generation_config_kwargs", invalid_generation_configs)
 @pytest.mark.parametrize("model_id", get_chat_models_list())
 @pytest.mark.parametrize("pipeline_type", [PipelineType.CONTINUOUS_BATCHING, PipelineType.SPECULATIVE_DECODING, PipelineType.PROMPT_LOOKUP_DECODING,])
-@pytest.mark.precommit
 def test_continuous_batching_add_request_fails(model_id, generation_config_kwargs: dict, pipeline_type):
     _, _, models_path = download_and_convert_model(model_id)
 
@@ -228,7 +222,6 @@ def test_continuous_batching_add_request_fails(model_id, generation_config_kwarg
 #
 
 # todo: iefode: bug reproducer!!!
-@pytest.mark.precommit
 @pytest.mark.parametrize("sampling_config", [get_greedy(), get_beam_search(), get_multinomial_all_parameters()],
                          ids=["greedy", "beam_search", "multinomial_all_parameters"])
 def test_post_oom_health(sampling_config):
@@ -289,7 +282,6 @@ def get_beam_search_seq_len_300() -> GenerationConfig:
                          ({"num_kv_blocks": 100, "dynamic_split_fuse": True}, get_beam_search_seq_len_300()),
                          ({"num_kv_blocks": 100, "dynamic_split_fuse": False}, get_beam_search_seq_len_300())]
 @pytest.mark.parametrize("params", scheduler_params_list)
-@pytest.mark.precommit
 def test_preemption(params):
     model_id = "facebook/opt-125m"
     scheduler_params = params[0]
@@ -342,7 +334,6 @@ def test_preemption(params):
 
 # todo: Anastasiia Pnevskaya: fix the test because it is hanging according max_new_tokens = std::numeric_limits<std::size_t>::max()
 @pytest.mark.parametrize("dynamic_split_fuse", [True, False])
-@pytest.mark.precommit
 @pytest.mark.skip(reason="Random sampling results are non deterministic due to: discrete_distribution impl depends on platform, model inference results may depend on CPU. Test passes on CI but fails locally.")
 def test_preemption_with_multinomial(dynamic_split_fuse):
     generation_configs = multinomial_params.generation_config
@@ -425,7 +416,6 @@ def test_preemption_with_multinomial(dynamic_split_fuse):
 
 
 @pytest.mark.parametrize("dynamic_split_fuse", [True, False])
-@pytest.mark.precommit
 @pytest.mark.skip(reason="Random sampling results are non deterministic due to: discrete_distribution impl depends on platform, model inference results may depend on CPU. Test passes on CI but fails locally.")
 def test_preemption_with_multinomial_n_seq(dynamic_split_fuse):
     model_id : str = "facebook/opt-125m"
@@ -442,7 +432,6 @@ def test_preemption_with_multinomial_n_seq(dynamic_split_fuse):
 
 
 @pytest.mark.parametrize("pipeline_type", [PipelineType.PROMPT_LOOKUP_DECODING])
-@pytest.mark.precommit
 def test_dynamic_split_fuse_doesnt_affect_generated_text(pipeline_type):
     model_id : str = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
     _, _, models_path = download_and_convert_model(model_id)
@@ -496,7 +485,6 @@ def run_extended_perf_metrics_collection(model_id, generation_config: Generation
 
 
 @pytest.mark.parametrize("pipeline_type", [PipelineType.PAGED_ATTENTION, PipelineType.SPECULATIVE_DECODING])
-@pytest.mark.precommit
 def test_speculative_decoding_extended_perf_metrics(pipeline_type):
     import time
     start_time = time.perf_counter()
diff --git a/tests/python_tests/test_generation_config.py b/tests/python_tests/test_generation_config.py
index 4acaa37648..2a5b1fa28a 100644
--- a/tests/python_tests/test_generation_config.py
+++ b/tests/python_tests/test_generation_config.py
@@ -61,7 +61,6 @@ def verify_set_values(generation_config, kwargs):
     dict(max_new_tokens=1, apply_chat_template=False),
 ]
 @pytest.mark.parametrize("generation_config_kwargs", configs)
-@pytest.mark.precommit
 def test_valid_configs(generation_config_kwargs):
     config = GenerationConfig(**generation_config_kwargs)
     verify_set_values(config, generation_config_kwargs)
@@ -107,7 +106,6 @@ def test_valid_configs(generation_config_kwargs):
     # TODO: add tests for invalid properties
 ]
 @pytest.mark.parametrize("generation_config_kwargs", invalid_configs)
-@pytest.mark.precommit
 def test_invalid_generation_configs_throws(generation_config_kwargs):
     config = GenerationConfig(**generation_config_kwargs)
     with pytest.raises(RuntimeError):
@@ -123,7 +121,6 @@ def test_invalid_generation_configs_throws(generation_config_kwargs):
     dict(eos_token_id=1), # 'stop_token_ids' does not contain 'eos_token_id'
     dict(eos_token_id=1, stop_token_ids={2}), # 'stop_token_ids' is not empty, but does not contain 'eos_token_id'
 ])
-@pytest.mark.precommit
 def test_invalid_fields_assinment_rises(fields):
     config = GenerationConfig()
     for key, val in fields.items():
@@ -147,7 +144,6 @@ def load_genai_generation_config_from_file(configs: list[tuple], temp_path):
 
     return ov_generation_config
 
-@pytest.mark.precommit
 def test_multiple_eos_are_read_as_stop_token_ids(tmp_path):
     generation_config_json = {
         "eos_token_id": [
diff --git a/tests/python_tests/test_gguf_reader.py b/tests/python_tests/test_gguf_reader.py
index 63a69694d0..f91855773c 100644
--- a/tests/python_tests/test_gguf_reader.py
+++ b/tests/python_tests/test_gguf_reader.py
@@ -18,7 +18,6 @@
 
 @pytest.mark.parametrize("pipeline_type", get_gguf_pipeline_types())
 @pytest.mark.parametrize("model_ids", get_gguf_model_list())
-@pytest.mark.precommit
 @pytest.mark.skipif(sys.platform == "win32", reason="CVS-174065")
 def test_pipelines_with_gguf_generate(pipeline_type, model_ids):
     if sys.platform == 'darwin':
@@ -69,7 +68,6 @@ def test_pipelines_with_gguf_generate(pipeline_type, model_ids):
     '<|endoftext|><|endoftext|><|im_end|>', 
     '<|endoftext|> Why the Sky is Blue? <|im_end|>',
 ])
-@pytest.mark.precommit
 @pytest.mark.skipif(sys.platform == "win32", reason="CVS-174065")
 def test_full_gguf_pipeline(pipeline_type, model_ids, enable_save_ov_model, prompt):
     if sys.platform == 'darwin':
@@ -129,7 +127,6 @@ def test_full_gguf_pipeline(pipeline_type, model_ids, enable_save_ov_model, prom
 @pytest.mark.parametrize("pipeline_type", get_gguf_pipeline_types())
 @pytest.mark.parametrize("model_ids", [{"gguf_model_id": "Qwen/Qwen3-0.6B-GGUF", "gguf_filename": "Qwen3-0.6B-Q8_0.gguf"}])
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 172335")
-@pytest.mark.precommit
 @pytest.mark.skipif(sys.platform == "win32", reason="CVS-174065")
 def test_full_gguf_qwen3_pipeline(pipeline_type, model_ids):
     # Temporal testing solution until transformers starts to support qwen3 in GGUF format
diff --git a/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_1.py b/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_1.py
index 2e96c4a096..7f2b3a60eb 100644
--- a/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_1.py
+++ b/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_1.py
@@ -42,7 +42,6 @@ class CacheOptTestStruct:
 SHORT_CACHE_EVICTION_CONFIG = CacheEvictionConfig(start_size=32, recent_size=32, max_cache_size=96, aggregation_mode=AggregationMode.NORM_SUM)
 LONGBENCH_CACHE_EVICTION_CONFIG = CacheEvictionConfig(start_size=32, recent_size=128, max_cache_size=672, aggregation_mode=AggregationMode.NORM_SUM)
 
-@pytest.mark.precommit
 @pytest.mark.skipif(
     sys.platform in ("win32", "darwin"),
     reason=(
@@ -164,7 +163,6 @@ def get_beam_search_seq_len_300() -> GenerationConfig:
                          ({"num_kv_blocks": 0, "cache_size": 0, "dynamic_split_fuse": False, "max_num_batched_tokens": 600, "enable_prefix_caching": False}, get_beam_search_seq_len_300()),
                          ({"num_kv_blocks": 0, "cache_size": 0, "dynamic_split_fuse": False, "max_num_batched_tokens": 600, "use_cache_eviction": True, "cache_eviction_config": SHORT_CACHE_EVICTION_CONFIG}, get_greedy_seq_len_300())]
 @pytest.mark.parametrize("params", scheduler_params_list)
-@pytest.mark.precommit
 def test_dynamic_memory_allocation(params):
     prompts, _ = get_test_dataset()
     generate_and_compare(prompts=prompts,
@@ -182,7 +180,6 @@ class LongBenchTestData:
     avg_cache_usage_optimization_ratio: float
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("test_struct", [
     LongBenchTestData("samsum", 4, 1.6, 2.5),
     LongBenchTestData("trec", 3.2, 2.0, 3.3),
diff --git a/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_2.py b/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_2.py
index 7ab1a98f91..1b9240c901 100644
--- a/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_2.py
+++ b/tests/python_tests/test_kv_cache_eviction/test_kv_cache_eviction_2.py
@@ -33,7 +33,6 @@
 }
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("subset", ["samsum", "trec", "qasper"])
 def test_kvcrush_vs_snapkv_baseline_longbench(subset):
     """Test that KVCrush performs equal or better than SnapKV baseline on LongBench datasets."""
diff --git a/tests/python_tests/test_llm_pipeline.py b/tests/python_tests/test_llm_pipeline.py
index 36b4688bcf..282fc07d1d 100644
--- a/tests/python_tests/test_llm_pipeline.py
+++ b/tests/python_tests/test_llm_pipeline.py
@@ -40,7 +40,6 @@ def assert_hf_equals_genai(hf_reference, genai_output):
 @pytest.mark.parametrize("generation_config_dict,prompt", test_cases)
 @pytest.mark.parametrize("model_id", get_models_list())
 @pytest.mark.parametrize("pipeline_type", get_main_pipeline_types())
-@pytest.mark.precommit
 def test_string_inputs(model_id, generation_config_dict, prompt, pipeline_type):
     generate_and_compare(model=model_id, prompts=[prompt], generation_config=generation_config_dict, pipeline_type=pipeline_type)
 
@@ -52,7 +51,6 @@ def test_string_inputs(model_id, generation_config_dict, prompt, pipeline_type):
 ]
 @pytest.mark.parametrize("inputs", input_tensors_list)
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_encoded_inputs(model_id, inputs):
     opt_model, hf_tokenizer, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -92,12 +90,10 @@ def test_encoded_inputs(model_id, inputs):
 @pytest.mark.parametrize("prompts", batched_prompts)
 @pytest.mark.parametrize("model_id", get_models_list())
 @pytest.mark.parametrize("pipeline_type", get_main_pipeline_types())
-@pytest.mark.precommit
 def test_batch_string_inputs(model_id, generation_config_dict, prompts, pipeline_type):
     generate_and_compare(model=model_id, prompts=prompts, generation_config=generation_config_dict, pipeline_type=pipeline_type)
 
 
-@pytest.mark.precommit
 def test_batch_size_switch():
     model_id = 'katuni4ka/tiny-random-phi3'
     _, _, models_path = download_and_convert_model(model_id)
@@ -108,7 +104,6 @@ def test_batch_size_switch():
     ov_pipe.generate(["a"], max_new_tokens=2)
 
 
-@pytest.mark.precommit
 def test_empty_encoded_inputs_throw():
     model_id = 'katuni4ka/tiny-random-phi3'
     _, _, models_path = download_and_convert_model(model_id)
@@ -118,7 +113,6 @@ def test_empty_encoded_inputs_throw():
         ov_pipe.generate(ov.Tensor(np.array([[]], dtype=np.int64)), max_new_tokens=2)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_chat_models_list())
 def test_different_input_types_works_same_and_change_nothing(model_id):
     opt_model, hf_tokenizer, models_path  = download_and_convert_model(model_id)
@@ -164,7 +158,6 @@ def test_different_input_types_works_same_and_change_nothing(model_id):
     GenerationChatInputsType.STRING,
     GenerationChatInputsType.ENCODED_INPUTS,
     GenerationChatInputsType.CHAT_HISTORY])
-@pytest.mark.precommit
 def test_chat_scenario(model_id, inputs, input_type):
     chat_history_hf = []
     chat_history_ov = ov_genai.ChatHistory() if input_type == GenerationChatInputsType.CHAT_HISTORY else []
@@ -236,7 +229,6 @@ def test_chat_scenario(model_id, inputs, input_type):
         assert_hf_equals_genai(chat_history_hf, chat_history_messages_ov)
 
 
-@pytest.mark.precommit
 def test_chat_scenario_several_chats_in_series():
     opt_model, hf_tokenizer, models_path  = download_and_convert_model(get_chat_models_list()[0])
     ov_pipe = create_ov_pipeline(models_path)
@@ -269,7 +261,6 @@ def test_chat_scenario_several_chats_in_series():
         assert_hf_equals_genai(chat_history_hf, chat_history_ov)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_chat_models_list())
 def test_chat_scenario_several_start(model_id):
     opt_model, hf_tokenizer, models_path  = download_and_convert_model(model_id)
@@ -284,7 +275,6 @@ def test_chat_scenario_several_start(model_id):
     ov_pipe.finish_chat()
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_chat_models_list())
 def test_generate_works_same_before_and_after_chat(model_id):
     opt_model, hf_tokenizer, models_path  = download_and_convert_model(model_id)
@@ -319,7 +309,6 @@ def user_defined_status_callback(subword):
 
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_callback_one_string(callback, model_id):
     _, _, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -330,7 +319,6 @@ def test_callback_one_string(callback, model_id):
 
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_callback_batch_throws(callback, model_id):
     _, _, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -340,7 +328,6 @@ def test_callback_batch_throws(callback, model_id):
 
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_callback_kwargs_one_string(callback, model_id):
     _, _, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -349,7 +336,6 @@ def test_callback_kwargs_one_string(callback, model_id):
 
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_callback_decoding_metallama(model_id, callback):
     # On metallama this prompt generates output which can shorten after adding new tokens.
     # Test that streamer correctly handles such cases.
@@ -363,7 +349,6 @@ def test_callback_decoding_metallama(model_id, callback):
 
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_callback_kwargs_batch_throws(callback, model_id):
     _, _, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -371,7 +356,6 @@ def test_callback_kwargs_batch_throws(callback, model_id):
         ov_pipe.generate(['1', '2'], max_new_tokens=10, streamer=callback)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_callback_terminate_by_bool(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -395,7 +379,6 @@ def callback(subword):
     assert len(ov_output.tokens[0]) < max_new_tokens
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_callback_terminate_by_status(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -420,7 +403,6 @@ def callback(subword):
 
 
 @pytest.mark.parametrize("model_id", get_chat_models_list())
-@pytest.mark.precommit
 def test_chat_scenario_callback_cancel(model_id):
     callback_questions = [
         '1+1=',
@@ -517,7 +499,6 @@ def end(self):
 
 @pytest.mark.parametrize("streamer_base", [PrinterNone, PrinterBool, PrinterStatus])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_streamer_one_string(streamer_base, model_id):
     _, _, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -527,7 +508,6 @@ def test_streamer_one_string(streamer_base, model_id):
     ov_pipe.generate('table is made of', generation_config, printer)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_streamer_batch_throws(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -537,7 +517,6 @@ def test_streamer_batch_throws(model_id):
         ov_pipe.generate(['1', '2'], ov_pipe.get_generation_config(), printer)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_streamer_kwargs_one_string(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -546,7 +525,6 @@ def test_streamer_kwargs_one_string(model_id):
     ov_pipe.generate('table is made of', max_new_tokens=10, do_sample=False, streamer=printer)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_streamer_kwargs_batch_throws(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -556,7 +534,6 @@ def test_streamer_kwargs_batch_throws(model_id):
         ov_pipe.generate('', num_beams=2, streamer=printer)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_operator_with_callback_one_string(callback, model_id):
@@ -567,7 +544,6 @@ def test_operator_with_callback_one_string(callback, model_id):
     ov_pipe('talbe is made of', ten_tokens, callback)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("callback", [print, user_defined_callback, user_defined_status_callback, lambda subword: print(subword)])
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_operator_with_callback_batch_throws(callback, model_id):
@@ -579,7 +555,6 @@ def test_operator_with_callback_batch_throws(callback, model_id):
 
 @pytest.mark.parametrize("streamer_base", [PrinterNone, PrinterBool, PrinterStatus])
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_operator_with_streamer_kwargs_one_string(streamer_base, model_id):
     _, _, models_path = download_and_convert_model(model_id)
     ov_pipe = create_ov_pipeline(models_path)
@@ -587,7 +562,6 @@ def test_operator_with_streamer_kwargs_one_string(streamer_base, model_id):
     ov_pipe('hi', max_new_tokens=10, do_sample=True, streamer=printer)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_operator_with_streamer_kwargs_batch_throws(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -620,7 +594,6 @@ def load_genai_pipe_with_configs(configs: list[tuple], temp_path):
     return ov_pipe
 
 
-@pytest.mark.precommit
 def test_eos_token_is_inherited_from_default_generation_config(model_tmp_path):
     _, temp_path = model_tmp_path
     ov_pipe = load_genai_pipe_with_configs([({"eos_token_id": 37}, "config.json")], temp_path)
@@ -632,7 +605,6 @@ def test_eos_token_is_inherited_from_default_generation_config(model_tmp_path):
     assert 37 == ov_pipe.get_generation_config().eos_token_id
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_pipeline_validates_generation_config(model_id):
     _, _, models_path = download_and_convert_model(model_id)
@@ -645,7 +617,6 @@ def test_pipeline_validates_generation_config(model_id):
 # Work with Unicode in Python API
 #
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_unicode_pybind_decoding_one_string(model_id):
     # On this model this prompt generates unfinished utf string.
@@ -656,7 +627,6 @@ def test_unicode_pybind_decoding_one_string(model_id):
     assert '�' == res_str[-1]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_unicode_pybind_decoding_batched(model_id):
     # On this model this prompt generates unfinished utf string.
@@ -667,7 +637,6 @@ def test_unicode_pybind_decoding_batched(model_id):
     assert '�' == res_str.texts[0][-1]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_unicode_pybind_decoding_one_string_streamer(model_id):
     # On this model this prompt generates unfinished utf-8 string
@@ -692,7 +661,6 @@ def run_perf_metrics_collection(model_id, generation_config_dict: dict, prompt:
     (dict(max_new_tokens=20), 'table is made of'),
 ]
 @pytest.mark.parametrize("generation_config,prompt", test_cases)
-@pytest.mark.precommit
 def test_perf_metrics(generation_config, prompt):
     import time
     start_time = time.perf_counter()
@@ -769,7 +737,6 @@ def test_perf_metrics(generation_config, prompt):
     (dict(max_new_tokens=20), 'Generate json of a person'),
 ]
 @pytest.mark.parametrize("generation_config,prompt", test_cases)
-@pytest.mark.precommit
 def test_perf_metrics_with_structured_output(generation_config, prompt):
     class Person(BaseModel):
         name: str = Field(pattern=r"^[A-Z][a-z]{1,20}$")
@@ -804,7 +771,6 @@ class Person(BaseModel):
 
 @pytest.mark.parametrize("pipeline_type", get_main_pipeline_types())
 @pytest.mark.parametrize("stop_str", {True, False})
-@pytest.mark.precommit
 def test_pipelines_generate_with_streaming(pipeline_type, stop_str):
     # streamer
     it_cnt = 0
diff --git a/tests/python_tests/test_llm_pipeline_static.py b/tests/python_tests/test_llm_pipeline_static.py
index e6a7cdc623..772cc5c8bd 100644
--- a/tests/python_tests/test_llm_pipeline_static.py
+++ b/tests/python_tests/test_llm_pipeline_static.py
@@ -65,7 +65,6 @@ def generate_with_chat_history(model_path, device, pipeline_config, questions) -
     get_greedy(),
     get_greedy_with_penalties()
 ]
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config", generation_configs)
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
@@ -83,7 +82,6 @@ def test_generation_compare_with_stateful(generation_config, config, model_id, i
     assert ref_out.texts[0] == actual_out.texts[0]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("with_weights", blob_with_weights)
 @pytest.mark.parametrize("model_id", get_models_list())
@@ -118,7 +116,6 @@ def test_pipeline_from_blob(model_tmp_path, config, with_weights, model_id):
     assert ref_out == actual_out
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("with_weights", blob_with_weights)
 @pytest.mark.parametrize("model_id", get_models_list())
@@ -162,7 +159,6 @@ def test_pipeline_cache_dir(model_tmp_path, config, with_weights, model_id):
 generation_configs = [
     get_multinomial_temperature_and_presence_penalty()
 ]
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config", generation_configs)
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
@@ -179,7 +175,6 @@ def test_multinomial_sampling(generation_config, config, model_id):
     actual_out = static_pipe.generate(prompt, generation_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_length_properties_set_no_exception(config, model_id):
@@ -199,7 +194,6 @@ def test_length_properties_set_no_exception(config, model_id):
 @pytest.mark.parametrize("length_config", length_configs)
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_invalid_length_properties_raise_error(length_config, config, model_id):
     _, _, model_path = download_and_convert_model(model_id)
     length_config |= config
@@ -207,7 +201,6 @@ def test_invalid_length_properties_raise_error(length_config, config, model_id):
         pipe = LLMPipeline(model_path, "NPU", **length_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_batch_one_no_exception(config, model_id):
@@ -219,7 +212,6 @@ def test_batch_one_no_exception(config, model_id):
 
 
 # TODO: For the further batch support
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_batch_raise_error(config, model_id):
@@ -239,7 +231,6 @@ def test_batch_raise_error(config, model_id):
 @pytest.mark.parametrize("generation_config", generation_configs)
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_unsupported_sampling_raise_error(generation_config, config, model_id):
     _, _, model_path = download_and_convert_model(model_id)
     prompt = 'What is OpenVINO?'
@@ -249,7 +240,6 @@ def test_unsupported_sampling_raise_error(generation_config, config, model_id):
         pipe.generate(prompt, generation_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_terminate_by_max_number_of_tokens(config, model_id):
@@ -265,7 +255,6 @@ def test_terminate_by_max_number_of_tokens(config, model_id):
     assert len(encoded_results.tokens[0]) == num_tokens
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_terminate_by_out_of_memory(config, model_id):
@@ -285,7 +274,6 @@ def test_terminate_by_out_of_memory(config, model_id):
     assert len(encoded_results.tokens[0]) == (kv_cache_size - input_len + 1)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
 def test_terminate_by_sampler(config, model_id):
@@ -318,7 +306,6 @@ def end(self):
 @pytest.mark.skip(reason="JIRA-144780: Output differs from stateful pipeline")
 @pytest.mark.parametrize("config", pipeline_configs)
 @pytest.mark.parametrize("model_id", get_models_list())
-@pytest.mark.precommit
 def test_chat_generation(config, model_id):
     questions = [
         '1+1=',
diff --git a/tests/python_tests/test_parsers.py b/tests/python_tests/test_parsers.py
index 3155e17fe1..1282b63b24 100644
--- a/tests/python_tests/test_parsers.py
+++ b/tests/python_tests/test_parsers.py
@@ -23,7 +23,6 @@ def hf_ov_genai_models(request, tmp_path_factory):
     return hf_tokenizer, genai_tokenizer
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     ["katuni4ka/tiny-random-phi3"],  # this tokenizer is used as a stub only
@@ -63,7 +62,6 @@ def write(self, message):
     assert msg['content'] == content
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     ["katuni4ka/tiny-random-phi3"],  # this tokenizer is used as a stub only
@@ -92,7 +90,6 @@ def write(self, message):
     assert msg['content'] == content
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     ["katuni4ka/tiny-random-phi3"],  # this tokenizer is used as a stub only
@@ -145,7 +142,6 @@ def write(self, message):
     assert msg['content'] == " The answer to 2 + 1 is 3."
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     ["katuni4ka/tiny-random-phi3"],
@@ -180,7 +176,6 @@ def write(self, message):
     assert msg['content'] == content
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("answer", [
     "<think>\nOkay, the user is asking for the answer to 2 + 1.</think>\n\nThe answer to 2 + 1 is \boxed{3}.",
 ])
@@ -203,7 +198,6 @@ def test_incremental_phi4_reason_parser_nostreamer(answer):
     assert msg['content'] == content
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("keep_original_content", [True, False])
 @pytest.mark.parametrize("do_reset", [False])
 @pytest.mark.parametrize(
@@ -270,7 +264,6 @@ def test_incremental_deepseek_parser():
     assert msg['content'] == content
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     ["katuni4ka/tiny-random-phi3"],
@@ -313,7 +306,6 @@ def write(self, message):
     assert msg['main_text'] == " world "
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     ["katuni4ka/tiny-random-phi3"],
diff --git a/tests/python_tests/test_rag.py b/tests/python_tests/test_rag.py
index 3facb41112..47b87fb259 100644
--- a/tests/python_tests/test_rag.py
+++ b/tests/python_tests/test_rag.py
@@ -290,7 +290,6 @@ def run_text_rerank_pipeline_with_ref(
 
 
 @pytest.mark.parametrize("download_and_convert_embeddings_models", ["BAAI/bge-small-en-v1.5"], indirect=True)
-@pytest.mark.precommit
 def test_embedding_constructors(download_and_convert_embeddings_models):
     _, _, models_path = download_and_convert_embeddings_models
 
@@ -325,7 +324,6 @@ def test_embedding_constructors(download_and_convert_embeddings_models):
         TextEmbeddingPipeline.Config(normalize=False, pooling_type=TextEmbeddingPipeline.PoolingType.LAST_TOKEN),
     ],
 )
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 174635")
 def test_qwen3_embedding(download_and_convert_embeddings_models, dataset_documents, config):
     opt_model, hf_tokenizer, models_path = download_and_convert_embeddings_models
@@ -355,7 +353,6 @@ def test_qwen3_embedding(download_and_convert_embeddings_models, dataset_documen
         "embed_instruction",
     ],
 )
-@pytest.mark.precommit
 def test_embed_documents(download_and_convert_embeddings_models, dataset_documents, config):
     if (
         sys.platform == "linux"
@@ -389,7 +386,6 @@ def test_embed_documents(download_and_convert_embeddings_models, dataset_documen
         "query_instruction",
     ],
 )
-@pytest.mark.precommit
 def test_embed_query(download_and_convert_embeddings_models, dataset_documents, config):
     _, _, models_path = download_and_convert_embeddings_models
     run_text_embedding_pipeline_with_ref(models_path, dataset_documents[:1], config, "embed_query")
@@ -414,7 +410,6 @@ def dataset_embeddings_genai_default_config_refs(download_and_convert_embeddings
         TextEmbeddingPipeline.Config(max_length=64, pad_to_max_length=True, batch_size=1),
     ],
 )
-@pytest.mark.precommit
 def test_fixed_shapes_configs(download_and_convert_embeddings_models, dataset_documents, config, dataset_embeddings_genai_default_config_refs):
     _, _, models_path = download_and_convert_embeddings_models
 
@@ -438,7 +433,6 @@ def test_fixed_shapes_configs(download_and_convert_embeddings_models, dataset_do
     ],
 )
 @pytest.mark.xfail()
-@pytest.mark.precommit
 def test_fixed_shapes_configs_xfail(download_and_convert_embeddings_models, dataset_documents, config, dataset_embeddings_genai_default_config_refs):
     _, _, models_path = download_and_convert_embeddings_models
 
@@ -457,7 +451,6 @@ def test_fixed_shapes_configs_xfail(download_and_convert_embeddings_models, data
         TextEmbeddingPipeline.Config(max_length=50, pad_to_max_length=True, batch_size=4),
     ],
 )
-@pytest.mark.precommit
 @pytest.mark.skipif(
     sys.platform == "darwin" or platform.machine() in ["aarch64", "arm64", "ARM64"],
     reason="NPU plugin is available only on Linux and Windows x86_64",
@@ -476,7 +469,6 @@ def test_npu_fallback(download_and_convert_embeddings_models, dataset_documents,
 
 
 @pytest.mark.parametrize("download_and_convert_rerank_model", [RERANK_TEST_MODELS[0]], indirect=True)
-@pytest.mark.precommit
 def test_rerank_constructors(download_and_convert_rerank_model):
     _, _, models_path = download_and_convert_rerank_model
 
@@ -514,7 +506,6 @@ def test_rerank_constructors(download_and_convert_rerank_model):
         "top_n=10",
     ],
 )
-@pytest.mark.precommit
 def test_rerank_documents(download_and_convert_rerank_model, dataset_documents, query, config):
     _, _, models_path = download_and_convert_rerank_model
     run_text_rerank_pipeline_with_ref(models_path, query, dataset_documents, config)
@@ -544,7 +535,6 @@ def test_rerank_documents(download_and_convert_rerank_model, dataset_documents,
         "top_n=4",
     ],
 )
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 174635")
 def test_qwen3_seq_cls_rerank_documents(download_and_convert_rerank_model, query, task, documents, config):
     opt_model, hf_tokenizer, models_path = download_and_convert_rerank_model
@@ -581,7 +571,6 @@ def test_qwen3_seq_cls_rerank_documents(download_and_convert_rerank_model, query
         "top_n=4",
     ],
 )
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 174635")
 def test_qwen3_rerank_documents(download_and_convert_model_fixture, query, task, documents, config):
     opt_model, hf_tokenizer, models_path = download_and_convert_model_fixture
diff --git a/tests/python_tests/test_sampling.py b/tests/python_tests/test_sampling.py
index fbd296bfcb..8e73a36bb4 100644
--- a/tests/python_tests/test_sampling.py
+++ b/tests/python_tests/test_sampling.py
@@ -12,7 +12,6 @@
 from utils.ov_genai_pipelines import generate_and_compare, run_ov_pipeline, get_main_pipeline_types
 from utils.hugging_face import download_and_convert_model
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config,prompt",
                          [(dict(max_new_tokens=30), 'table is made of'),
                           (dict(max_new_tokens=30, min_new_tokens=30), '你好！ 你好嗎？'),
@@ -33,7 +32,6 @@ def test_basic_stop_criteria(generation_config, prompt):
     generate_and_compare(model_id, [prompt], generation_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config,model_id",
                          [(dict(max_new_tokens=50, min_new_tokens=15, stop_strings={"anag"}, include_stop_str_in_output=True), 'facebook/opt-125m'), # expected match on "manage"
                           (dict(max_new_tokens=50, min_new_tokens=1, stop_strings={".", "software", "Intel"}, include_stop_str_in_output=True), 'facebook/opt-125m'),
@@ -57,7 +55,6 @@ def test_stop_strings(generation_config, model_id, pipeline_type):
     generate_and_compare(model_id, prompts, generation_config, pipeline_type=pipeline_type)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config",
                          [dict(max_new_tokens=30),
                           dict(max_new_tokens=30, repetition_penalty=2.0),
@@ -79,7 +76,6 @@ def test_greedy(generation_config, prompt):
                          generation_config=generation_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config",
                          [dict(max_new_tokens=30, num_beams=2),
                           dict(max_new_tokens=30, num_beams=2, stop_criteria=StopCriteria.NEVER),
@@ -105,7 +101,6 @@ def test_beam_search(generation_config):
     generate_and_compare(model_id, prompts, generation_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.xfail(
     raises=AssertionError,
     reason="Stop strings do not seem to work as expected with beam search in HF, so comparison will fail. If it changes, these cases shall be merged to the test above.",
@@ -121,7 +116,6 @@ def test_beam_search_with_stop_string(generation_config):
     generate_and_compare(model_id, prompts, generation_config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("generation_config",
                          [dict(max_new_tokens=1, min_new_tokens=0, echo=True),
                           dict(max_new_tokens=30, num_beams=2, echo=True),],
@@ -312,7 +306,6 @@ class RandomSamplingTestStruct:
 ]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("test_struct", RANDOM_SAMPLING_TEST_CASES,
         ids=["multinomial_temperature",
              "multinomial_temperature_and_top_p",
diff --git a/tests/python_tests/test_stateful_speculative_decoding.py b/tests/python_tests/test_stateful_speculative_decoding.py
index 9aa90836e9..f4be11f328 100644
--- a/tests/python_tests/test_stateful_speculative_decoding.py
+++ b/tests/python_tests/test_stateful_speculative_decoding.py
@@ -29,7 +29,6 @@ def get_npu_llm_properties_for_test():
 ]
 @pytest.mark.parametrize("main_model,draft_model,prompt", models_and_input)
 @pytest.mark.parametrize("main_device,draft_device", devices)
-@pytest.mark.precommit
 def test_string_inputs(main_model, main_device, draft_model, draft_device, prompt):
     # Download and convert model:
     main_opt_model, main_hf_tokenizer, main_model_path = download_and_convert_model(main_model)
@@ -67,7 +66,6 @@ def test_string_inputs(main_model, main_device, draft_model, draft_device, promp
     compare_generation_results([prompt], ref_gen_results, ov_chat_history_gen_results, ov_generation_config)
 
 
-@pytest.mark.precommit
 def test_perf_metrics():
     import time
     start_time = time.perf_counter()
@@ -147,7 +145,6 @@ def test_perf_metrics():
     assert len(raw_metrics.m_batch_sizes) > 0
     assert len(raw_metrics.m_durations) > 0
 
-@pytest.mark.precommit
 def test_extended_perf_metrics():
     import time
     start_time = time.perf_counter()
diff --git a/tests/python_tests/test_structured_output.py b/tests/python_tests/test_structured_output.py
index 98b704095e..9cc1b74691 100644
--- a/tests/python_tests/test_structured_output.py
+++ b/tests/python_tests/test_structured_output.py
@@ -39,7 +39,6 @@ class RESTAPIResponse(BaseModel):
 ]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("ov_pipe", structured_id_models, indirect=True)
 @pytest.mark.parametrize(
     "prompt_and_scheme",
@@ -70,7 +69,6 @@ def test_structured_json(ov_pipe, prompt_and_scheme, use_compound_grammar, capfd
         pytest.fail(f"Output {res_str} is not valid json schema {SchemeType.model_json_schema()}: {e}")
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("ov_pipe", structured_id_models, indirect=True)
 @pytest.mark.parametrize(
     "prompt_and_regex",
@@ -98,7 +96,6 @@ def test_structured_regex(ov_pipe, prompt_and_regex, use_compound_grammar):
     assert re.match(regex_str, res_str), f"Output {res_str} does not match regex {regex_str}"
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("ov_pipe", structured_id_models, indirect=True)
 @pytest.mark.parametrize(
     "prompt_and_ebnf",
@@ -138,7 +135,6 @@ def test_structured_ebnf(ov_pipe, prompt_and_ebnf, use_compound_grammar):
     assert re.match(r"^\d{4}-\d{2}-\d{2}$", res_str), f"Output {res_str} does not match date format"
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "ov_pipe", [model_id for model_id in structured_id_models if "random" not in model_id], indirect=True
 )
@@ -173,7 +169,6 @@ def test_structural_tags_old(ov_pipe, prompt_and_structural_tag):
     RESTAPIResponse.model_validate_json(match.group(1))
 
 
-@pytest.mark.precommit
 # use only non-random model for stable output in TriggeredTags test
 @pytest.mark.parametrize("ov_pipe", ["TinyLlama/TinyLlama-1.1B-Chat-v1.0"], indirect=True)
 @pytest.mark.parametrize(
diff --git a/tests/python_tests/test_text_streamer.py b/tests/python_tests/test_text_streamer.py
index 75804256b1..a8df543da9 100644
--- a/tests/python_tests/test_text_streamer.py
+++ b/tests/python_tests/test_text_streamer.py
@@ -52,7 +52,6 @@ def chunks(arr: list, n: int):
 ])]
 
 @pytest.mark.parametrize("model_id", tokenizer_model_ids)
-@pytest.mark.precommit
 @pytest.mark.parametrize("prompt", [*eng_prompts, *unicode_prompts])
 def test_text_prompts(tmp_path, prompt, model_id):
     prompt = prompt.decode('unicode_escape') if isinstance(prompt, bytes) else prompt
@@ -93,7 +92,6 @@ def test_text_prompts(tmp_path, prompt, model_id):
     [167, 96, 227, 169, 232, 250, 167, 96, 227, 169, 232, 250, 167]
 ]
 @pytest.mark.parametrize("model_id", tokenizer_model_ids)
-@pytest.mark.precommit
 @pytest.mark.parametrize("encoded_prompt", encoded_prompts)
 def test_encoded_prompts(tmp_path, encoded_prompt, model_id):
     model_id, hf_tok_load_params = (model_id[0], model_id[1]) if isinstance(model_id, tuple) else (model_id, {})
diff --git a/tests/python_tests/test_tokenizer.py b/tests/python_tests/test_tokenizer.py
index eeba864d5a..0b6065af3c 100644
--- a/tests/python_tests/test_tokenizer.py
+++ b/tests/python_tests/test_tokenizer.py
@@ -70,7 +70,6 @@ def ov_hf_tokenizers(request):
 
 @pytest.mark.parametrize("ov_hf_tokenizers", get_models_list(), indirect=True)
 @pytest.mark.parametrize("prompt", prompts)
-@pytest.mark.precommit
 def test_encode(ov_hf_tokenizers, prompt):
     ov_tokenizer, hf_tokenizer = ov_hf_tokenizers
 
@@ -102,7 +101,6 @@ def test_encode(ov_hf_tokenizers, prompt):
 
 @pytest.mark.parametrize("ov_hf_tokenizers", get_models_list(), indirect=True)
 @pytest.mark.parametrize("encoded_prompt", encoded_prompts)
-@pytest.mark.precommit
 def test_decode(ov_hf_tokenizers, encoded_prompt):
     ov_tokenizer, hf_tokenizer = ov_hf_tokenizers
     decoded_ov = ov_tokenizer.decode(encoded_prompt)
@@ -130,7 +128,6 @@ def test_decode(ov_hf_tokenizers, encoded_prompt):
 ]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("chat_config", get_chat_templates())
 @pytest.mark.parametrize("ov_hf_tokenizers", get_models_list(), indirect=True)
 def test_apply_chat_template(model_tmp_path, chat_config: tuple[str, dict], ov_hf_tokenizers):
@@ -152,7 +149,6 @@ def test_apply_chat_template(model_tmp_path, chat_config: tuple[str, dict], ov_h
     assert_hf_equals_genai(hf_full_history_str, ov_full_history_str)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("ov_hf_tokenizers", get_models_list(), indirect=True)
 @pytest.mark.parametrize("tokenizer_config_model_id", ["google/gemma-3-1b-it"])
 def test_apply_chat_template_nested_content(model_tmp_path, ov_hf_tokenizers, tokenizer_config_model_id):
@@ -192,7 +188,6 @@ def test_apply_chat_template_nested_content(model_tmp_path, ov_hf_tokenizers, to
     assert genai_templated_chat_history == ov_full_history_str
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("ov_hf_tokenizers", get_models_list(), indirect=True)
 @pytest.mark.parametrize("tokenizer_config_model_id", ["Qwen/Qwen3-8B-Base"])
 def test_apply_chat_template_with_tools_and_extra_context(model_tmp_path, ov_hf_tokenizers, tokenizer_config_model_id):
@@ -247,7 +242,6 @@ def test_apply_chat_template_with_tools_and_extra_context(model_tmp_path, ov_hf_
     assert_hf_equals_genai(genai_templated_chat_history, ov_full_history_str)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "hf_ov_genai_models", 
     [("Xenova/c4ai-command-r-v01-tokenizer", { "padding_side": None })],
@@ -265,7 +259,6 @@ def test_non_string_chat_template(hf_ov_genai_models):
     assert_hf_equals_genai(hf_full_history_str, ov_full_history_str)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("ov_hf_tokenizers", get_models_list(), indirect=True)
 def test_set_chat_template(ov_hf_tokenizers):
     ov_tokenizer, hf_tokenizer = ov_hf_tokenizers
@@ -314,7 +307,6 @@ def test_set_chat_template(ov_hf_tokenizers):
     ],
     indirect=True,
 )
-@pytest.mark.precommit
 @pytest.mark.parametrize("prompt", [*eng_prompts, *unicode_prompts])
 def test_special_tokens(prompt, ov_hf_tokenizers):
     prompt = prompt.decode("unicode_escape") if isinstance(prompt, bytes) else prompt
@@ -346,7 +338,6 @@ def test_special_tokens(prompt, ov_hf_tokenizers):
     assert decoded_hf_skip_spec != decoded_hf_no_skip
 
 
-@pytest.mark.precommit
 def test_multiple_infer_request_state(tmp_path):
     hf_tokenizer = retry_request(lambda: AutoTokenizer.from_pretrained("llamafactory/tiny-random-Llama-3"))
     ov_tokenizer = convert_tokenizer(hf_tokenizer)
@@ -410,7 +401,6 @@ def hf_ov_genai_models(request, tmp_path_factory):
 ]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("add_special_tokens", [True, False])
 @pytest.mark.parametrize("max_length", [None, 16, 103, 512, 1024])
 @pytest.mark.parametrize("pad_to_max_length", [None, True, False])
@@ -506,7 +496,6 @@ def make_model_params():
 models_with_pair_input = make_model_params()
 
 @pytest.mark.parametrize("hf_ov_genai_models", models_with_pair_input, indirect=True)
-@pytest.mark.precommit
 @pytest.mark.parametrize("input_pair", [[
     ["hi", "sun in yellow"],
     ["Eng... test, string?!" * 100, "Multiline\nstring!\nWow!"],
@@ -522,7 +511,6 @@ def test_two_inputs_string_list_of_lists_batched(hf_ov_genai_models, input_pair)
     assert np.all(ov_encoded == hf_encoded)
 
 @pytest.mark.parametrize("hf_ov_genai_models", models_with_pair_input, indirect=True)
-@pytest.mark.precommit
 @pytest.mark.parametrize("input_pair", [
     [["hi", "sun in yellow"]],
     [["Eng... test, string?!" * 100, "Multiline\nstring!\nWow!"]],
@@ -539,7 +527,6 @@ def test_two_inputs_string_list_of_lists(hf_ov_genai_models, input_pair):
 
 
 @pytest.mark.parametrize("hf_ov_genai_models", models_with_pair_input, indirect=True)
-@pytest.mark.precommit
 @pytest.mark.parametrize("input_pair", [
     [["Eng... test, string?!" * 100], ["Multiline\nstring!\nWow!"]],
     [["hi" * 20], ["buy" * 90]],
@@ -562,7 +549,6 @@ def test_two_inputs_string(hf_ov_genai_models, input_pair):
     assert np.all(ov_encoded == hf_encoded)
 
 
-@pytest.mark.precommit
 def test_load_special_tokens_from_config_json(model_tmp_path):
     # test when there is an available config.json
     config_json = {
@@ -576,7 +562,6 @@ def test_load_special_tokens_from_config_json(model_tmp_path):
     assert tok.get_eos_token_id() == config_json["eos_token_id"]
 
 
-@pytest.mark.precommit
 def test_load_special_tokens_from_special_tokens_map_json(model_tmp_path):
     # test with special_tokens_map
     special_tokens_map_json = {
@@ -590,7 +575,6 @@ def test_load_special_tokens_from_special_tokens_map_json(model_tmp_path):
     assert tok.get_eos_token() == special_tokens_map_json["eos_token"]["content"]
 
 
-@pytest.mark.precommit
 def test_load_special_tokens_from_tokenizer_config_json(model_tmp_path):
     # special_tokens_map is not available
     # but tokenize_config.json exists
@@ -616,7 +600,6 @@ def test_load_special_tokens_from_tokenizer_config_json(model_tmp_path):
     assert tok.get_eos_token_id() == 42
 
 
-@pytest.mark.precommit
 def test_load_special_tokens_from_tokenizer_config_and_config_json(model_tmp_path):
     # both config.json is available and tokenizer_config.json available
     # check that it does not read int values from tokenizer_config.json if they are in config.json
@@ -647,7 +630,6 @@ def test_load_special_tokens_from_tokenizer_config_and_config_json(model_tmp_pat
     assert tok.get_eos_token() == tok_config_json["eos_token"]
 
 
-@pytest.mark.precommit
 @pytest.mark.xfail(
     raises=AssertionError,
     reason="CVS-143410 ov tokenizer should be aligned with hf",
@@ -721,14 +703,12 @@ def generate_tokenizer(tmp_path, chat_templates):
 SIMPLIFIED_QWEN2_VL_2B = "{% for message in messages %}{{ message['content'] }}{% endfor %}"
 
 
-@pytest.mark.precommit
 def test_set_special_runtime_template(tmp_path):
     tokenizer = generate_tokenizer(tmp_path, ChatTemplates(None, None, None, None, None))
     tokenizer.chat_template = QWEN2_VL_2B
     assert tokenizer.chat_template == SIMPLIFIED_QWEN2_VL_2B
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "chat_templates",
     [
@@ -745,7 +725,6 @@ def test_template_priorities(tmp_path, chat_templates):
     assert tokenizer.chat_template == chat_templates.reference
 
 
-@pytest.mark.precommit
 def test_chat_template_with_empty_output(tmp_path):
     tokenizer = generate_tokenizer(tmp_path, ChatTemplates(None, None, None, None, None))
     # Test throwing exception for empty rendered chat template (e.g. Qwen2-VL)
diff --git a/tests/python_tests/test_vlm_pipeline.py b/tests/python_tests/test_vlm_pipeline.py
index d72051c35c..c9a567ce16 100644
--- a/tests/python_tests/test_vlm_pipeline.py
+++ b/tests/python_tests/test_vlm_pipeline.py
@@ -420,7 +420,6 @@ def test_images(request: pytest.FixtureRequest):
     return [request.getfixturevalue(image) for image in request.param]
 
 
-@pytest.mark.precommit
 @parametrize_all_models
 def test_vlm_pipeline(ov_pipe_model: VlmModelInfo, test_images: list[openvino.Tensor]):
     ov_pipe = ov_pipe_model.pipeline
@@ -442,7 +441,6 @@ def streamer(word: str) -> bool:
     assert res.texts[0] == "".join(result_from_streamer)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "config", 
     [
@@ -506,7 +504,6 @@ def test_vlm_continuous_batching_generate_vs_add_request(
             )
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "config", 
     [
@@ -618,7 +615,6 @@ def iteration_images_and_videos(request):
     return params
 
 
-@pytest.mark.precommit
 @parametrize_all_models
 @pytest.mark.parametrize("system_message", ["", "You are a helpful assistant."])
 def test_vlm_pipeline_chat(
@@ -679,7 +675,6 @@ def iteration_images_npu(request):
     return [[request.getfixturevalue(image) for image in bundle] for bundle in request.param]
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", MODEL_IDS)
 @pytest.mark.parametrize("system_message", ["", "You are a helpful assistant."])
 def test_vlm_pipeline_chat_npu(model_id, system_message, iteration_images_npu):
@@ -721,7 +716,6 @@ def streamer(word: str) -> bool:
     run_chat(npu_pipe, system_message, iteration_images_npu)
 
 
-@pytest.mark.precommit
 @parametrize_all_models_with_video
 @pytest.mark.parametrize("system_message", ["", "You are a helpful assistant."])
 def test_vlm_pipeline_chat_with_video(
@@ -771,7 +765,6 @@ def streamer(word: str) -> bool:
     ov_pipe.finish_chat()
 
 
-@pytest.mark.precommit
 @parametrize_one_model_backends
 def test_vlm_get_tokenizer(ov_pipe_model: VlmModelInfo):
     ov_pipe = ov_pipe_model.pipeline
@@ -779,7 +772,6 @@ def test_vlm_get_tokenizer(ov_pipe_model: VlmModelInfo):
     tokenizer.encode("")
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "config",
     [
@@ -797,7 +789,6 @@ def test_sampling(
     ov_pipe.generate(PROMPTS[0], image=cat_tensor, generation_config=config)
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("backend", ATTENTION_BACKEND)
 def test_perf_metrics(
     backend: str, 
@@ -865,7 +856,6 @@ def test_perf_metrics(
     assert np.allclose(std_dur, np.std(raw_dur))
 
 
-@pytest.mark.precommit
 @pytest.mark.parametrize("model_id", MODEL_IDS)
 @pytest.mark.parametrize("backend", ATTENTION_BACKEND)
 @pytest.mark.skipif(
@@ -901,7 +891,6 @@ def image_sequence(request):
     return [request.getfixturevalue(image) for image in request.param]
 
 
-@pytest.mark.precommit
 @pytest.mark.skipif(
     sys.platform == "darwin" or platform.machine() in ["aarch64", "arm64", "ARM64"],
     reason="NPU plugin is available only on Linux and Windows x86_64",
@@ -923,7 +912,6 @@ def test_vlm_npu_no_image():
     )
 
 
-@pytest.mark.precommit
 @parametrize_all_models
 def test_vlm_pipeline_chat_streamer_cancel_second_generate(
     request: pytest.FixtureRequest,
@@ -998,7 +986,6 @@ def streamer(subword):
     assert results_with_cancel == results
 
 
-@pytest.mark.precommit
 @parametrize_one_model_backends
 def test_start_chat_clears_history(
     ov_pipe_model: VlmModelInfo,
@@ -1025,7 +1012,6 @@ def test_start_chat_clears_history(
     assert results_first_generate == results_second_generate
 
 
-@pytest.mark.precommit
 def test_start_chat_clears_history_cb_api(
     ov_continious_batching_pipe: ContinuousBatchingPipeline, 
     image_sequence: list[openvino.Tensor]
@@ -1051,7 +1037,6 @@ def test_start_chat_clears_history_cb_api(
     assert results_first_generate == results_second_generate
 
 
-@pytest.mark.precommit
 @parametrize_all_models
 def test_vlm_pipeline_chat_streamer_cancel_first_generate(
     request: pytest.FixtureRequest,
@@ -1182,7 +1167,6 @@ def model_and_tag_parametrize(
     )
 
 
-@pytest.mark.precommit
 @model_and_tag_parametrize(TAG_INSERTED_BY_TEMPLATE)
 def test_model_tags_representation(ov_pipe_model: VlmModelInfo, cat_tensor: openvino.Tensor):
     ov_pipe = ov_pipe_model.pipeline
@@ -1235,7 +1219,6 @@ def workaround_inconsistent_inference():
     retry(workaround_inconsistent_inference)
 
 
-@pytest.mark.precommit
 @model_and_tag_parametrize()
 def test_model_tags_prepend_native(
     ov_pipe_model: VlmModelInfo, 
@@ -1264,7 +1247,6 @@ def workaround_inconsistent_inference():
     retry(workaround_inconsistent_inference)
 
 
-@pytest.mark.precommit
 @model_and_tag_parametrize()
 def test_model_tags_prepend_universal(
     ov_pipe_model: VlmModelInfo, 
@@ -1297,7 +1279,6 @@ def workaround_inconsistent_inference():
 def cat_image_384x384(cat_image):
     return cat_image.resize((384, 384))
 
-@pytest.mark.precommit
 @model_and_tag_parametrize()
 def test_model_tags_append(
     ov_pipe_model: VlmModelInfo, 
@@ -1338,7 +1319,6 @@ def workaround_inconsistent_inference():
     retry(workaround_inconsistent_inference)
 
 
-@pytest.mark.precommit
 @model_and_tag_parametrize(IMAGE_ID_IGNORANT_MODELS_TO_TAG)
 def test_model_tags_same_reference(ov_pipe_model: VlmModelInfo, cat_tensor: openvino.Tensor):
     ov_pipe = ov_pipe_model.pipeline
@@ -1358,7 +1338,6 @@ def workaround_inconsistent_inference():
     retry(workaround_inconsistent_inference)
 
 
-@pytest.mark.precommit
 @model_and_tag_parametrize()
 def test_model_tags_older(ov_pipe_model: VlmModelInfo, car_tensor: openvino.Tensor):
     ov_pipe = ov_pipe_model.pipeline
@@ -1372,7 +1351,6 @@ def test_model_tags_older(ov_pipe_model: VlmModelInfo, car_tensor: openvino.Tens
     ov_pipe.finish_chat()
         
         
-@pytest.mark.precommit
 @model_and_tag_parametrize()
 def test_model_tags_missing_universal(ov_pipe_model: VlmModelInfo):
     ov_pipe = ov_pipe_model.pipeline
@@ -1381,7 +1359,6 @@ def test_model_tags_missing_universal(ov_pipe_model: VlmModelInfo):
         ov_pipe.generate("<ov_genai_image_0>")
         
         
-@pytest.mark.precommit
 @model_and_tag_parametrize()
 def test_model_tags_missing_native(ov_pipe_model: VlmModelInfo):
     ov_pipe = ov_pipe_model.pipeline
@@ -1391,7 +1368,6 @@ def test_model_tags_missing_native(ov_pipe_model: VlmModelInfo):
         ov_pipe.generate(image_tag(0))
             
 
-@pytest.mark.precommit
 @pytest.mark.parametrize(
     "ov_pipe_model,has_image,has_video",
     [
diff --git a/tests/python_tests/test_whisper_pipeline.py b/tests/python_tests/test_whisper_pipeline.py
index 655f527852..66c08335a9 100644
--- a/tests/python_tests/test_whisper_pipeline.py
+++ b/tests/python_tests/test_whisper_pipeline.py
@@ -275,7 +275,6 @@ def compare_results(hf_result, genai_result):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language": "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_smoke(model_descr, sample_from_dataset):
     run_pipeline_with_ref(
         model_id=model_descr[0],
@@ -285,7 +284,6 @@ def test_smoke(model_descr, sample_from_dataset):
 
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
-@pytest.mark.precommit
 def test_whisper_config_constructor(model_descr):
     model_id, path = model_descr
 
@@ -325,7 +323,6 @@ def test_whisper_config_constructor(model_descr):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language" : "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_whisper_constructors(model_descr, sample_from_dataset):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
 
@@ -345,7 +342,6 @@ def test_whisper_constructors(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_max_new_tokens(model_descr, sample_from_dataset):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
 
@@ -363,7 +359,6 @@ def test_max_new_tokens(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("language", ["fr", "de"])
-@pytest.mark.precommit
 def test_language_mode(model_descr, language):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
     sample = get_whisper_dataset(language, long_form=False)[0]
@@ -388,7 +383,6 @@ def test_language_mode(model_descr, language):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", get_fixture_params_for_n_whisper_dataset_samples(n=1, language="fr"), indirect=True)
-@pytest.mark.precommit
 def test_task_mode(model_descr, sample_from_dataset):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
 
@@ -438,7 +432,6 @@ def test_task_mode(model_descr, sample_from_dataset):
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=1, language="fr"),
                                                  *get_fixture_params_for_n_whisper_dataset_samples(n=1, language="de"),
                                                  *get_fixture_params_for_n_whisper_dataset_samples(n=1, language="es")], indirect=True)
-@pytest.mark.precommit
 def test_language_autodetect(model_descr, sample_from_dataset):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
 
@@ -457,7 +450,6 @@ def test_language_autodetect(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=1)], indirect=True)
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_return_timestamps_short_form(model_descr, sample_from_dataset):
     run_pipeline_with_ref(
@@ -470,7 +462,6 @@ def test_return_timestamps_short_form(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=1)], indirect=True)
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_return_timestamps_max_new_tokens_short_form(model_descr, sample_from_dataset):
     run_pipeline_with_ref(
@@ -485,7 +476,6 @@ def test_return_timestamps_max_new_tokens_short_form(model_descr, sample_from_da
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list())
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=10, long_form=True)], indirect=True)
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_longform_audio(model_descr, sample_from_dataset):
     _, _, hf_pipe, genai_pipe = read_whisper_model(model_descr)
@@ -512,7 +502,6 @@ def test_longform_audio(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list())
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=2, long_form=True)], indirect=True)
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_longform_audio_with_past(model_descr, sample_from_dataset):
     _, _, hf_pipe, genai_pipe = read_whisper_model(model_descr, stateful=True)
@@ -538,7 +527,6 @@ def test_longform_audio_with_past(model_descr, sample_from_dataset):
 
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list())
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_shortform(model_descr):
     samples = []
@@ -558,7 +546,6 @@ def test_shortform(model_descr):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=2, long_form=True)], indirect=True)
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_beam_search(model_descr, sample_from_dataset):
     # use only 30 seconds of audio due to beam search results wrong with enabled timestamps
@@ -576,7 +563,6 @@ def test_beam_search(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language" : "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_initial_prompt_hotwords(model_descr, sample_from_dataset):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
 
@@ -598,7 +584,6 @@ def test_initial_prompt_hotwords(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language" : "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_random_sampling(model_descr, sample_from_dataset):
     _, _, hf_pipe, genai_pipe = read_whisper_model(model_descr)
 
@@ -637,7 +622,6 @@ def test_random_sampling(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language" : "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 @pytest.mark.xfail(condition=(sys.platform == "darwin"), reason="Ticket - 173169")
 def test_perf_metrics(model_descr, sample_from_dataset):
     model_id, path, hf_pipe, genai_pipe = read_whisper_model(model_descr)
@@ -785,7 +769,6 @@ def streamer_bool_callback(subword):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language" : "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_streamers(model_descr, sample_from_dataset, streamer_for_test):
     _, _, _, genai_pipe = read_whisper_model(model_descr)
 
diff --git a/tests/python_tests/test_whisper_pipeline_static.py b/tests/python_tests/test_whisper_pipeline_static.py
index 86e22bae60..08afe6215d 100644
--- a/tests/python_tests/test_whisper_pipeline_static.py
+++ b/tests/python_tests/test_whisper_pipeline_static.py
@@ -76,7 +76,6 @@ def compare_results_with_assert(expected, actual_out):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language": "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_generation_compare_with_cpu(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr)
 
@@ -89,7 +88,6 @@ def test_static_whisper_generation_compare_with_cpu(model_descr, sample_from_dat
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=2, language="fr"),
                                                  *get_fixture_params_for_n_whisper_dataset_samples(n=2, language="de"),
                                                  *get_fixture_params_for_n_whisper_dataset_samples(n=2, language="es")], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_autodetect(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr)
 
@@ -100,7 +98,6 @@ def test_static_whisper_autodetect(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(language='de', n=3)], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_language_de(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr)
 
@@ -111,7 +108,6 @@ def test_static_whisper_language_de(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(language='fr', n=3)], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_language_fr(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr)
 
@@ -122,7 +118,6 @@ def test_static_whisper_language_fr(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(language='ru', n=3)], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_language_ru(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr)
 
@@ -133,7 +128,6 @@ def test_static_whisper_language_ru(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language": "en", "sample_id": 0, "long_form": True}], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_generation_long(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr)
 
@@ -144,7 +138,6 @@ def test_static_whisper_generation_long(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language": "en", "sample_id": 0}], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_stateful_generation_compare_with_cpu(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr, stateful=True)
 
@@ -157,7 +150,6 @@ def test_static_whisper_stateful_generation_compare_with_cpu(model_descr, sample
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(n=2, language="fr"),
                                                  *get_fixture_params_for_n_whisper_dataset_samples(n=2, language="de"),
                                                  *get_fixture_params_for_n_whisper_dataset_samples(n=2, language="es")], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_stateful_autodetect(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr, stateful=True)
 
@@ -168,7 +160,6 @@ def test_static_whisper_stateful_autodetect(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(language='de', n=3)], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_stateful_language_de(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr, stateful=True)
 
@@ -179,7 +170,6 @@ def test_static_whisper_stateful_language_de(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(language='fr', n=3)], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_stateful_language_fr(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr, stateful=True)
 
@@ -190,7 +180,6 @@ def test_static_whisper_stateful_language_fr(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [*get_fixture_params_for_n_whisper_dataset_samples(language='ru', n=3)], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_stateful_language_ru(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr, stateful=True)
 
@@ -201,7 +190,6 @@ def test_static_whisper_stateful_language_ru(model_descr, sample_from_dataset):
 
 @pytest.mark.parametrize("model_descr", get_whisper_models_list(tiny_only=True))
 @pytest.mark.parametrize("sample_from_dataset", [{"language": "en", "sample_id": 0, "long_form": True}], indirect=True)
-@pytest.mark.precommit
 def test_static_whisper_stateful_generation_long(model_descr, sample_from_dataset):
     model_id, model_path = load_and_save_whisper_model(model_descr, stateful=True)