diff --git a/python/sglang/srt/arg_groups/overrides.py b/python/sglang/srt/arg_groups/overrides.py
index 07329b6cd..6a516be2b 100644
--- a/python/sglang/srt/arg_groups/overrides.py
+++ b/python/sglang/srt/arg_groups/overrides.py
@@ -31,6 +31,7 @@ from __future__ import annotations
 
 import dataclasses
 import inspect
+import json
 import logging
 from typing import Any, Callable, Dict, List, Optional, Sequence, Tuple
 
@@ -290,6 +291,42 @@ def attention_backends_of(cfg: Any) -> tuple:
     return prefill, decode
 
 
+def modelexpress_transport_of(cfg: Any) -> str:
+    """The modelexpress transport a config-shaped object asks for.
+
+    ``modelexpress_config`` is a JSON string (or an already-parsed dict) rather
+    than a leaf of its own; this is the shared parse for the transfer-engine
+    gate (`remote_instance_transfer_engine_of`) and any future bag reader.
+    ``ServerArgs.modelexpress_transport`` keeps its own instance-cached parse
+    (`_parsed_modelexpress_config`) -- same rule, cached seed-side."""
+    raw = cfg.modelexpress_config
+    if raw is None:
+        parsed = {}
+    elif isinstance(raw, str):
+        parsed = json.loads(raw)
+    else:
+        parsed = raw
+    return parsed.get("transport", "nixl")
+
+
+def remote_instance_transfer_engine_of(cfg: Any, load_format: Any = None) -> bool:
+    """Whether remote-instance weight loading runs over the transfer engine.
+
+    ``load_format`` overrides the config's: a draft runner loading under
+    ``--speculative-draft-load-format`` needs its own transfer engine. Every
+    input is a ``model`` leaf, so this serves both the pre-publish member and
+    the post-publish accessor."""
+    if cfg.remote_instance_weight_loader_start_seed_via_transfer_engine:
+        return True
+    if (load_format or cfg.load_format) != "remote_instance":
+        return False
+    backend = cfg.remote_instance_weight_loader_backend
+    return backend == "transfer_engine" or (
+        backend == "modelexpress"
+        and modelexpress_transport_of(cfg) == "transfer_engine"
+    )
+
+
 def mamba_extra_buffer_of(cfg: Any) -> bool:
     """Mid-resolution equivalent of runtime_context.mamba_extra_buffer_enabled:
     reads the (possibly overlaid) strategy from a config-shaped object.
diff --git a/python/sglang/srt/managers/scheduler.py b/python/sglang/srt/managers/scheduler.py
index 09b0a9421..eb6041cd3 100644
--- a/python/sglang/srt/managers/scheduler.py
+++ b/python/sglang/srt/managers/scheduler.py
@@ -1876,7 +1876,7 @@ class Scheduler(
     def process_input_requests(self, recv_reqs: List):
         now = time.monotonic()
         self.session_controller.maybe_reap(now)
-        if self.server_args.mm_feature_transport == "cuda_vmm":
+        if get_mm().mm_feature_transport == "cuda_vmm":
             for recv_req in recv_reqs:
                 self._materialize_cuda_vmm_inputs(recv_req)
 
diff --git a/python/sglang/srt/mem_cache/kv_cache_configurator.py b/python/sglang/srt/mem_cache/kv_cache_configurator.py
index 6e36069e5..1b4dfd78c 100644
--- a/python/sglang/srt/mem_cache/kv_cache_configurator.py
+++ b/python/sglang/srt/mem_cache/kv_cache_configurator.py
@@ -75,6 +75,7 @@ from sglang.srt.runtime_context import (
     mamba_extra_buffer_enabled,
     mamba_extra_buffer_lazy_enabled,
     max_speculative_num_draft_tokens,
+    pre_capture_activation_reserve_mb,
 )
 from sglang.srt.server_args import ServerArgs
 from sglang.srt.speculative.spec_info import SpeculativeAlgorithm
@@ -1782,7 +1783,7 @@ class KVCacheConfigurator:
             # Mamba state is a fixed pre-capture allocation, so it can't ride the ~0 post-capture slack.
             slack_gb = max(
                 slack_gb,
-                self.server_args.pre_capture_activation_reserve_mb(
+                pre_capture_activation_reserve_mb(
                     get_device_memory_capacity(self.device)
                 )
                 / 1024,
diff --git a/python/sglang/srt/model_executor/model_runner.py b/python/sglang/srt/model_executor/model_runner.py
index 7cf7c00ca..7f754bc24 100644
--- a/python/sglang/srt/model_executor/model_runner.py
+++ b/python/sglang/srt/model_executor/model_runner.py
@@ -179,6 +179,7 @@ from sglang.srt.runtime_context import (
     get_spec,
     is_ep_joiner,
     is_ep_scale_joiner,
+    remote_instance_transfer_engine_enabled,
     set_global_dwdp_manager,
 )
 from sglang.srt.sampling.sampling_batch_info import SamplingBatchInfo
@@ -560,7 +561,6 @@ class ModelRunner:
 
     def init_remote_instance_weight_transporter(self):
         self.remote_instance_weight_transporter = RemoteInstanceWeightTransporter(
-            server_args=self.server_args,
             get_model=lambda: self.model,
             tp_rank=self.ps.tp_rank,
             gpu_id=self.gpu_id,
@@ -671,9 +671,7 @@ class ModelRunner:
         )
 
     def maybe_init_remote_instance_transfer_engine(self):
-        if self.server_args.remote_instance_weight_loader_use_transfer_engine(
-            load_format=self.draft_load_format
-        ):
+        if remote_instance_transfer_engine_enabled(load_format=self.draft_load_format):
             self.remote_instance_weight_transporter.init_engine()
 
     def maybe_init_expert_location_metadata(self):
diff --git a/python/sglang/srt/model_executor/model_runner_components/kv_pool_runtime.py b/python/sglang/srt/model_executor/model_runner_components/kv_pool_runtime.py
index 58c315149..9be8dfb43 100644
--- a/python/sglang/srt/model_executor/model_runner_components/kv_pool_runtime.py
+++ b/python/sglang/srt/model_executor/model_runner_components/kv_pool_runtime.py
@@ -11,6 +11,7 @@ from sglang.srt.distributed import get_world_group
 from sglang.srt.mem_cache.kv_cache_configurator import mm_runtime_reservation_gb
 from sglang.srt.model_executor.cuda_graph_config import Backend
 from sglang.srt.platforms import current_platform
+from sglang.srt.runtime_context import pre_capture_activation_reserve_mb
 from sglang.srt.utils.common import get_available_gpu_memory, get_device_memory_capacity
 
 if TYPE_CHECKING:
@@ -72,7 +73,7 @@ def compute_post_capture_kv_resize(
     if eager_decode_gap or mambaish_config(model_runner.model_config) is not None:
         headroom_gb = max(
             headroom_gb,
-            model_runner.server_args.pre_capture_activation_reserve_mb(
+            pre_capture_activation_reserve_mb(
                 get_device_memory_capacity(model_runner.device)
             )
             / 1024,
diff --git a/python/sglang/srt/model_executor/model_runner_components/remote_instance_weight_transporter.py b/python/sglang/srt/model_executor/model_runner_components/remote_instance_weight_transporter.py
index 5fbcab266..14690ab21 100644
--- a/python/sglang/srt/model_executor/model_runner_components/remote_instance_weight_transporter.py
+++ b/python/sglang/srt/model_executor/model_runner_components/remote_instance_weight_transporter.py
@@ -11,8 +11,11 @@ from sglang.srt.model_loader.remote_instance_weight_loader_utils import (
     RemoteInstanceWeightLoaderBackend,
     register_memory_region,
 )
-from sglang.srt.runtime_context import get_model, get_parallel
-from sglang.srt.server_args import ServerArgs
+from sglang.srt.runtime_context import (
+    get_model,
+    get_parallel,
+    remote_instance_transfer_engine_enabled,
+)
 from sglang.srt.utils.network import NetworkAddress, get_local_ip_auto
 
 logger = logging.getLogger(__name__)
@@ -20,7 +23,6 @@ logger = logging.getLogger(__name__)
 
 @dataclass(slots=True, kw_only=True)
 class RemoteInstanceWeightTransporter:
-    server_args: ServerArgs
     get_model: Callable[[], torch.nn.Module]
     tp_rank: int
     gpu_id: int
@@ -55,7 +57,7 @@ class RemoteInstanceWeightTransporter:
 
     def maybe_register_and_publish_weight_info(self) -> None:
         if (
-            self.server_args.remote_instance_weight_loader_use_transfer_engine()
+            remote_instance_transfer_engine_enabled()
             # ModelExpress owns TransferEngine memory registration and metadata
             # publishing for backend=modelexpress. Re-registering here would
             # overlap the same weight buffers.
diff --git a/python/sglang/srt/runtime_context.py b/python/sglang/srt/runtime_context.py
index b572581bd..a71ebff22 100644
--- a/python/sglang/srt/runtime_context.py
+++ b/python/sglang/srt/runtime_context.py
@@ -1420,6 +1420,53 @@ def mamba_extra_buffer_lazy_enabled() -> bool:
     )
 
 
+def remote_instance_transfer_engine_enabled(load_format: str | None = None) -> bool:
+    """Whether remote-instance weight loading runs over the transfer engine.
+
+    Every input is a ``model`` leaf, so this derives from the bags and follows a
+    post-publish override; ``ServerArgs.remote_instance_weight_loader_use_transfer_engine``
+    is the pre-publish equivalent, and both go through the same helper.
+    ``load_format`` is the caller's own (a draft runner loading under
+    ``--speculative-draft-load-format`` has one the process record does not).
+    """
+    from sglang.srt.arg_groups.overrides import remote_instance_transfer_engine_of
+
+    return remote_instance_transfer_engine_of(get_model(), load_format)
+
+
+def pre_capture_activation_reserve_mb(gpu_mem: float | None) -> float:
+    """The activation working-set reserve held back before cuda-graph capture.
+
+    Derived from published leaves across four bags (``disagg`` / ``schedule`` /
+    ``exec.graph`` / ``spec``) plus the configured parallel sizes, so it follows
+    a post-publish override; ``ServerArgs.pre_capture_activation_reserve_mb`` is
+    the pre-publish equivalent and
+    ``TestDerivedPredicatesAgreeAcrossTiers`` pins the two equal.
+    """
+    schedule = get_schedule()
+    if get_disagg().disaggregation_mode == "decode":
+        running_requests = (
+            schedule.max_running_requests
+            or get_exec().graph.cuda_graph_config.decode.max_bs
+            or 1
+        )
+        activation_tokens = max(
+            running_requests * (get_spec().speculative_num_draft_tokens or 1), 2048
+        )
+    elif schedule.chunked_prefill_size > 0:
+        activation_tokens = max(schedule.chunked_prefill_size, 2048)
+    else:
+        activation_tokens = max(schedule.max_prefill_tokens, 2048)
+    reserved_mem = (
+        512
+        + activation_tokens * 1.5
+        + _configured_parallel("tp_size") * _configured_parallel("pp_size") / 8 * 1024
+    )
+    if gpu_mem is not None and gpu_mem > 60 * 1024:
+        reserved_mem = max(reserved_mem, 10 * 1024)
+    return reserved_mem
+
+
 # --- Derived config accessors ------------------------------------------------
 #
 # A few values are computed from several config fields plus the HF config, so
diff --git a/python/sglang/srt/server_args.py b/python/sglang/srt/server_args.py
index 2c2813eb7..2def4ccb2 100644
--- a/python/sglang/srt/server_args.py
+++ b/python/sglang/srt/server_args.py
@@ -44,6 +44,7 @@ from sglang.srt.arg_groups.overrides import (
     attention_backends_of,
     mamba_extra_buffer_lazy_of,
     mamba_extra_buffer_of,
+    remote_instance_transfer_engine_of,
     resolved_view,
 )
 from sglang.srt.configs.embedding_model_spec import BCGPrefillPolicy
@@ -9470,20 +9471,7 @@ class ServerArgs:
     def remote_instance_weight_loader_use_transfer_engine(self, load_format=None):
         """``load_format`` overrides the seed's: a draft runner loading under
         ``--speculative-draft-load-format`` needs its own transfer engine."""
-        # Use TransferEngine as seed backend.
-        if self.remote_instance_weight_loader_start_seed_via_transfer_engine:
-            return True
-        # Use TransferEngine as client backend.
-        if (load_format or self.load_format) == "remote_instance" and (
-            self.remote_instance_weight_loader_backend == "transfer_engine"
-            or (
-                self.remote_instance_weight_loader_backend == "modelexpress"
-                and self.modelexpress_transport == "transfer_engine"
-            )
-        ):
-            return True
-        else:
-            return False
+        return remote_instance_transfer_engine_of(self, load_format)
 
     def describe_kv_events_publisher(self) -> Optional[dict]:
         """Return a structured description of this server's KV-event
diff --git a/python/sglang/srt/speculative/base_spec_worker.py b/python/sglang/srt/speculative/base_spec_worker.py
index 259f9847f..6c2501515 100644
--- a/python/sglang/srt/speculative/base_spec_worker.py
+++ b/python/sglang/srt/speculative/base_spec_worker.py
@@ -11,7 +11,7 @@ from sglang.srt.model_executor.graph_memory_usage import (
     merge_graph_memory_usage,
     merge_graph_time_usage,
 )
-from sglang.srt.runtime_context import get_exec, get_schedule
+from sglang.srt.runtime_context import get_exec, get_memory, get_schedule
 
 if TYPE_CHECKING:
     from sglang.srt.managers.io_struct import (
@@ -235,7 +235,7 @@ class BaseSpecWorker(ABC):
         target_model_runner = self.target_worker.model_runner
         target_model_runner.mtp_draft_device_pools = ()
         spec_algorithm = target_model_runner.spec_algorithm
-        if not self.server_args.enable_hierarchical_cache:
+        if not get_memory().enable_hierarchical_cache:
             return HiCacheDraftPlan()
 
         draft_runners = self._draft_model_runners()
diff --git a/python/sglang/srt/speculative/draft_utils.py b/python/sglang/srt/speculative/draft_utils.py
index bf3e92f29..821d28874 100644
--- a/python/sglang/srt/speculative/draft_utils.py
+++ b/python/sglang/srt/speculative/draft_utils.py
@@ -1,5 +1,4 @@
-from sglang.srt.runtime_context import get_spec
-from sglang.srt.server_args import ServerArgs
+from sglang.srt.runtime_context import attention_backends, get_spec
 from sglang.srt.utils.common import (
     cpu_has_amx_support,
     is_blackwell,
@@ -28,13 +27,11 @@ def _assert_draft_needs_no_conv_sidecar(draft_model_runner) -> None:
 class DraftBackendFactory:
     def __init__(
         self,
-        server_args: ServerArgs,
         draft_model_runner,
         topk: int,
         speculative_num_steps: int,
         seed_dsa_topk_from_draft_extend: bool = False,
     ):
-        self.server_args = server_args
         self.draft_model_runner = draft_model_runner
         self.topk = topk
         self.speculative_num_steps = speculative_num_steps
@@ -45,13 +42,14 @@ class DraftBackendFactory:
     def _create_backend(
         self, backend_name: str, backend_map: dict, error_template: str
     ):
-        backend_type = (
-            self.draft_attn_backend
-            if self.draft_attn_backend
-            else getattr(self.server_args, backend_name)
+        # The split pair with the base-backend fallback already applied.
+        prefill_backend, decode_backend = attention_backends()
+        configured = (
+            decode_backend
+            if backend_name == "decode_attention_backend"
+            else prefill_backend
         )
-        if backend_type is None:
-            backend_type = self.server_args.attention_backend
+        backend_type = self.draft_attn_backend or configured
 
         if backend_type not in backend_map:
             raise ValueError(error_template.format(backend_type=backend_type))
diff --git a/python/sglang/srt/speculative/eagle_worker_v2.py b/python/sglang/srt/speculative/eagle_worker_v2.py
index 2c2fc649e..9fb14f970 100644
--- a/python/sglang/srt/speculative/eagle_worker_v2.py
+++ b/python/sglang/srt/speculative/eagle_worker_v2.py
@@ -320,7 +320,6 @@ class EagleDraftWorker(EagleDraftWorkerBase):
         self.draft_extend_attn_backend = None
 
         draft_backend_factory = DraftBackendFactory(
-            self.server_args,
             self.draft_runner,
             self.topk,
             self.speculative_num_steps,
diff --git a/python/sglang/srt/speculative/multi_layer_eagle_worker_v2.py b/python/sglang/srt/speculative/multi_layer_eagle_worker_v2.py
index b9280f8a2..bf449b9fd 100644
--- a/python/sglang/srt/speculative/multi_layer_eagle_worker_v2.py
+++ b/python/sglang/srt/speculative/multi_layer_eagle_worker_v2.py
@@ -362,7 +362,6 @@ class MultiLayerEagleDraftWorker(EagleDraftWorkerBase):
         self.draft_extend_attn_backend_list = []
         for step in range(self.speculative_num_steps):
             draft_backend_factory = DraftBackendFactory(
-                self.server_args,
                 self.draft_runner_list[step],
                 self.topk,
                 self.speculative_num_steps,
diff --git a/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_extend_runner.py b/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_extend_runner.py
index b3e7dd235..e0df0d4b2 100644
--- a/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_extend_runner.py
+++ b/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_extend_runner.py
@@ -498,7 +498,6 @@ def _build_eagle_draft_extend_fixture(
         speculative_attention_mode="prefill",
     )
     draft_extend_attn_backend = DraftBackendFactory(
-        fixture.runner.server_args,
         fixture.runner,
         settings.topk,
         settings.speculative_num_steps,
diff --git a/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_runner.py b/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_runner.py
index 964eec4c8..dd76e9187 100644
--- a/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_runner.py
+++ b/python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_runner.py
@@ -359,7 +359,6 @@ def _build_eagle_draft_fixture(
     )
     _configure_runner_for_eagle_draft(fixture.runner, case, settings)
     draft_attn_backend = DraftBackendFactory(
-        fixture.runner.server_args,
         fixture.runner,
         settings.topk,
         settings.speculative_num_steps,
diff --git a/test/registered/unit/multimodal/test_gpu_feature_transport.py b/test/registered/unit/multimodal/test_gpu_feature_transport.py
index 287869e61..c2e7e57c4 100644
--- a/test/registered/unit/multimodal/test_gpu_feature_transport.py
+++ b/test/registered/unit/multimodal/test_gpu_feature_transport.py
@@ -490,6 +490,13 @@ class TestCudaVmmFeatureTransport(unittest.TestCase):
 
 
 class TestSchedulerMmTransportBoundary(unittest.TestCase):
+    def _publish(self, **fields):
+        from sglang.srt.runtime_context import get_context
+
+        override = get_context().override_server_args(**fields)
+        override.install()
+        self.addCleanup(override.restore)
+
     @staticmethod
     def _prepare_scheduler(scheduler):
         scheduler.session_controller = SimpleNamespace(maybe_reap=MagicMock())
@@ -501,7 +508,9 @@ class TestSchedulerMmTransportBoundary(unittest.TestCase):
         from sglang.srt.managers import scheduler as scheduler_module
 
         scheduler = object.__new__(scheduler_module.Scheduler)
-        scheduler.server_args = SimpleNamespace(
+        # The transport gate reads the published bags, so the case publishes
+        # the configuration under test.
+        self._publish(
             mm_feature_transport="cuda_vmm",
             enable_broadcast_mm_inputs_process=True,
         )
@@ -548,7 +557,7 @@ class TestSchedulerMmTransportBoundary(unittest.TestCase):
                 return iter(self.batch)
 
         scheduler = object.__new__(scheduler_module.Scheduler)
-        scheduler.server_args = SimpleNamespace(mm_feature_transport="cuda_vmm")
+        self._publish(mm_feature_transport="cuda_vmm")
         self._prepare_scheduler(scheduler)
         raw_inputs = [object(), object()]
         materialized = [object(), object()]
diff --git a/test/registered/unit/test_runtime_context.py b/test/registered/unit/test_runtime_context.py
index 3d519b5cc..86e0f90af 100644
--- a/test/registered/unit/test_runtime_context.py
+++ b/test/registered/unit/test_runtime_context.py
@@ -391,6 +391,19 @@ class _FakeResolvedArgs:
     speculative_num_draft_tokens: A[int | None, Arg(help="d"), NS("spec")] = None
     speculative_adaptive: A[bool, Arg(help="a"), NS("spec")] = False
     speculative_adaptive_config: A[str | None, Arg(help="c"), NS("spec")] = None
+    load_format: A[str, Arg(help="lf"), NS("model")] = "auto"
+    remote_instance_weight_loader_backend: A[str, Arg(help="rb"), NS("model")] = "nccl"
+    remote_instance_weight_loader_start_seed_via_transfer_engine: A[
+        bool, Arg(help="rs"), NS("model")
+    ] = False
+    modelexpress_config: A[str | None, Arg(help="mx"), NS("model")] = None
+    disaggregation_mode: A[str, Arg(help="dm"), NS("disagg")] = "null"
+    max_running_requests: A[int | None, Arg(help="mrr"), NS("schedule")] = None
+    chunked_prefill_size: A[int, Arg(help="cps"), NS("schedule")] = -1
+    max_prefill_tokens: A[int, Arg(help="mpt"), NS("schedule")] = 16384
+    cuda_graph_config: A[object | None, Arg(help="cgc"), NS("exec.graph")] = None
+    tp_size: A[int, Arg(help="tp"), NS("parallel")] = 1
+    pp_size: A[int, Arg(help="pp"), NS("parallel")] = 1
     _resolved_overrides: list = dataclasses.field(default_factory=list)
 
 
@@ -1013,6 +1026,74 @@ class TestDerivedPredicatesAgreeAcrossTiers(_IsolatedServerArgs):
                         mamba_extra_buffer_lazy_enabled(),
                     )
 
+    def test_activation_reserve_matches_the_member(self):
+        from types import SimpleNamespace
+
+        from sglang.srt.runtime_context import pre_capture_activation_reserve_mb
+
+        graph = SimpleNamespace(decode=SimpleNamespace(max_bs=64))
+        cases = (
+            dict(disaggregation_mode="null", chunked_prefill_size=8192),
+            dict(disaggregation_mode="null", chunked_prefill_size=-1),
+            dict(
+                disaggregation_mode="null",
+                chunked_prefill_size=-1,
+                max_prefill_tokens=1024,
+            ),
+            dict(disaggregation_mode="decode", max_running_requests=32),
+            dict(disaggregation_mode="decode", max_running_requests=None),
+            dict(
+                disaggregation_mode="decode",
+                max_running_requests=None,
+                speculative_num_draft_tokens=4,
+            ),
+            dict(
+                disaggregation_mode="null",
+                chunked_prefill_size=8192,
+                tp_size=8,
+                pp_size=2,
+            ),
+        )
+        for case in cases:
+            for gpu_mem in (None, 20 * 1024, 80 * 1024):
+                with self.subTest(gpu_mem=gpu_mem, **case):
+                    args = _FakeResolvedArgs(cuda_graph_config=graph, **case)
+                    get_context().set_server_args(args)
+                    self.assertEqual(
+                        ServerArgs.pre_capture_activation_reserve_mb(args, gpu_mem),
+                        pre_capture_activation_reserve_mb(gpu_mem),
+                    )
+
+    def test_remote_instance_transfer_engine_matches_the_member(self):
+        from sglang.srt.runtime_context import remote_instance_transfer_engine_enabled
+
+        backends = ("nccl", "transfer_engine", "modelexpress")
+        transports = (None, '{"transport": "transfer_engine"}', '{"transport": "nixl"}')
+        for seed_via_te in (False, True):
+            for load_format in ("auto", "remote_instance"):
+                for backend in backends:
+                    for mx in transports:
+                        with self.subTest(
+                            seed=seed_via_te,
+                            load_format=load_format,
+                            backend=backend,
+                            modelexpress=mx,
+                        ):
+                            args = _FakeResolvedArgs(
+                                load_format=load_format,
+                                remote_instance_weight_loader_backend=backend,
+                                remote_instance_weight_loader_start_seed_via_transfer_engine=seed_via_te,
+                                modelexpress_config=mx,
+                            )
+                            get_context().set_server_args(args)
+                            for override in (None, "remote_instance", "auto"):
+                                self.assertEqual(
+                                    ServerArgs.remote_instance_weight_loader_use_transfer_engine(
+                                        args, override
+                                    ),
+                                    remote_instance_transfer_engine_enabled(override),
+                                )
+
     def test_attention_backends_match_the_member(self):
         from sglang.srt.runtime_context import attention_backends