diff --git a/test/srt/test_original_logprobs.py b/test/registered/sampling/test_original_logprobs.py
similarity index 97%
rename from test/srt/test_original_logprobs.py
rename to test/registered/sampling/test_original_logprobs.py
index 81ed0a403..f024a172e 100644
--- a/test/srt/test_original_logprobs.py
+++ b/test/registered/sampling/test_original_logprobs.py
@@ -14,6 +14,11 @@ The test covers the following scenarios:
 """
 
 import os
+
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+register_cuda_ci(est_time=41, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu")
 import random
 import unittest
 
diff --git a/test/srt/test_penalty.py b/test/registered/sampling/test_penalty.py
similarity index 97%
rename from test/srt/test_penalty.py
rename to test/registered/sampling/test_penalty.py
index 9c32e4ca4..a074dfaee 100644
--- a/test/srt/test_penalty.py
+++ b/test/registered/sampling/test_penalty.py
@@ -7,6 +7,10 @@ from concurrent.futures import ThreadPoolExecutor
 import requests
 
 from sglang.srt.utils import kill_process_tree
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+register_cuda_ci(est_time=82, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=180, suite="stage-b-test-small-1-gpu")
 from sglang.test.test_utils import (
     DEFAULT_SMALL_MODEL_NAME_FOR_TEST,
     DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
diff --git a/test/srt/test_pytorch_sampling_backend.py b/test/registered/sampling/test_pytorch_sampling_backend.py
similarity index 92%
rename from test/srt/test_pytorch_sampling_backend.py
rename to test/registered/sampling/test_pytorch_sampling_backend.py
index 8a368310e..058708b76 100644
--- a/test/srt/test_pytorch_sampling_backend.py
+++ b/test/registered/sampling/test_pytorch_sampling_backend.py
@@ -4,6 +4,10 @@ from types import SimpleNamespace
 import requests
 
 from sglang.srt.utils import kill_process_tree
+from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
+
+register_cuda_ci(est_time=66, suite="stage-b-test-small-1-gpu")
+register_amd_ci(est_time=66, suite="stage-b-test-small-1-gpu")
 from sglang.test.run_eval import run_eval
 from sglang.test.test_utils import (
     DEFAULT_MODEL_NAME_FOR_TEST,
diff --git a/test/srt/run_suite.py b/test/srt/run_suite.py
index a0d108f7c..9e990555b 100644
--- a/test/srt/run_suite.py
+++ b/test/srt/run_suite.py
@@ -48,12 +48,9 @@ suites = {
         TestFile("test_multi_tokenizer.py", 230),
         TestFile("test_no_chunked_prefill.py", 108),
         TestFile("test_no_overlap_scheduler.py", 217),
-        TestFile("test_original_logprobs.py", 41),
         TestFile("test_page_size.py", 60),
-        TestFile("test_penalty.py", 82),
         TestFile("test_prefill_adder.py", 1),
         TestFile("test_priority_scheduling.py", 130),
-        TestFile("test_pytorch_sampling_backend.py", 66),
         TestFile("test_request_queue_validation.py", 47),
         TestFile("test_retract_decode.py", 259),
         TestFile("test_score_api.py", 260),
@@ -189,14 +186,11 @@ suite_amd = {
         TestFile("test_model_hooks.py", 10),
         TestFile("test_multi_tokenizer.py", 345),
         TestFile("test_no_chunked_prefill.py", 108),
-        TestFile("test_original_logprobs.py", 60),
         TestFile("test_page_size.py", 60),
-        TestFile("test_penalty.py", 180),
         TestFile("test_prefill_adder.py", 2),
         TestFile("test_priority_scheduling.py", 195),
         TestFile("test_profile_merger.py", 12),
         TestFile("test_profile_merger_http_api.py", 15),
-        TestFile("test_pytorch_sampling_backend.py", 66),
         TestFile("test_request_queue_validation.py", 70),
         TestFile("test_retract_decode.py", 450),
         TestFile("test_rope_rocm.py", 3),