diff --git a/python/sglang/srt/managers/scheduler.py b/python/sglang/srt/managers/scheduler.py
index 50fd2e9e5..c4ee02cab 100644
--- a/python/sglang/srt/managers/scheduler.py
+++ b/python/sglang/srt/managers/scheduler.py
@@ -2016,7 +2016,7 @@ class Scheduler(
             draft_worker=self.draft_worker,
             model_worker=self.model_worker,
             logprob_result_processor=SchedulerLogprobResultProcessor(
-                server_args=self.server_args, model_config=self.model_config
+                model_config=self.model_config
             ),
             output_streamer=self.output_streamer,
             abort_request=self.abort_request,
diff --git a/python/sglang/srt/managers/scheduler_components/logprob_result_processor.py b/python/sglang/srt/managers/scheduler_components/logprob_result_processor.py
index f1f659565..379501f0e 100644
--- a/python/sglang/srt/managers/scheduler_components/logprob_result_processor.py
+++ b/python/sglang/srt/managers/scheduler_components/logprob_result_processor.py
@@ -14,17 +14,13 @@ from sglang.srt.layers.logits_processor import LogitsProcessorOutput
 from sglang.srt.managers.io_struct import build_flat_input_top_logprobs_arrays
 from sglang.srt.managers.schedule_batch import Req
 from sglang.srt.runtime_context import get_exec
-from sglang.srt.server_args import (
-    MIS_DELIMITER_TOKEN_ID,
-    ServerArgs,
-)
+from sglang.srt.server_args import MIS_DELIMITER_TOKEN_ID
 
 logger = logging.getLogger(__name__)
 
 
 @dataclass(kw_only=True, slots=True, frozen=True)
 class SchedulerLogprobResultProcessor:
-    server_args: ServerArgs
     model_config: ModelConfig
 
     def _process_input_token_logprobs(
diff --git a/test/registered/unit/managers/test_flat_raw_top_logprobs.py b/test/registered/unit/managers/test_flat_raw_top_logprobs.py
index d4225dd9a..a32e26d49 100644
--- a/test/registered/unit/managers/test_flat_raw_top_logprobs.py
+++ b/test/registered/unit/managers/test_flat_raw_top_logprobs.py
@@ -19,6 +19,7 @@ from sglang.test.test_utils import CustomTestCase, maybe_stub_sgl_kernel
 
 maybe_stub_sgl_kernel()
 
+from sglang.srt import runtime_context as rc
 from sglang.srt.managers.io_struct import (
     BatchTokenIDOutput,
     GenerateReqInput,
@@ -318,9 +319,8 @@ class TestB64MetaInfo(CustomTestCase):
 
 
 def _make_logprob_processor() -> SchedulerLogprobResultProcessor:
-    # The processor only reads enable_mis and vocab_size from these.
+    # enable_mis comes from the published exec bag, not from here; see setUp.
     return SchedulerLogprobResultProcessor(
-        server_args=SimpleNamespace(enable_mis=False),
         model_config=SimpleNamespace(vocab_size=1_000_000),
     )
 
@@ -341,6 +341,14 @@ _SCHED_IDX_ROWS = [[11, 22], [33, 44], [55, 66], [77, 88], [99, 100]]
 class TestSchedulerFlatAssembly(CustomTestCase):
     """Scheduler-side flat assembly in the logprob result processor."""
 
+    def setUp(self):
+        super().setUp()
+        self._server_args_override = rc.get_context().override_server_args()
+        self._server_args_override.install()
+
+    def tearDown(self):
+        self._server_args_override.restore()
+
     def _make_req(self, flat: bool, num_tokens: int = 5) -> Req:
         return Req(
             "r0",