diff --git a/Cargo.lock b/Cargo.lock
index 1808078eb21d..d7c9ea9a832a 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -34,9 +34,9 @@ dependencies = [
 
 [[package]]
 name = "aisimulate-core"
-version = "0.1.0-dev.2"
+version = "0.12.0-dev.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3aa0422f239fd9b9ca92d4a57fe7069771e158293d266e05a20fde477d15b9af"
+checksum = "d4db6a085073ecb1c2569f796ba44742ed4681f8f4060daaed2456ef4ea9b652"
 dependencies = [
  "anyhow",
  "bincode 1.3.3",
@@ -45,6 +45,7 @@ dependencies = [
  "derive_builder",
  "flate2",
  "indicatif 0.18.4",
+ "log",
  "parquet",
  "pep440_rs",
  "quick_cache",
diff --git a/Cargo.toml b/Cargo.toml
index ee9e06eb0945..e5bfe95b4187 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -56,7 +56,7 @@ keywords = ["llm", "genai", "inference", "nvidia", "distributed"]
 
 [workspace.dependencies]
 # Published separately by https://github.com/ai-dynamo/aisimulate.
-aisimulate-core = { version = "=0.1.0-dev.2" }
+aisimulate-core = { version = "=0.12.0-dev.1" }
 
 # Local crates
 dynamo-runtime = { path = "lib/runtime", version = "1.5.0" }
diff --git a/benchmarks/mocker/bench_aic_concurrency.py b/benchmarks/mocker/bench_aic_concurrency.py
index 05bb9687dd68..5cc2a8e92279 100644
--- a/benchmarks/mocker/bench_aic_concurrency.py
+++ b/benchmarks/mocker/bench_aic_concurrency.py
@@ -39,7 +39,7 @@
 MODEL = "Qwen/Qwen3-32B"
 SYSTEM = "h200_sxm"
 BACKEND = "vllm"
-BACKEND_VERSION = "0.19.0"
+BACKEND_VERSION = "current"
 
 BS, ISL = 16, 2048
 THREAD_COUNTS = (1, 2, 4, 8, 12, 16, 24)
diff --git a/benchmarks/mocker/bench_aic_rust_callback.py b/benchmarks/mocker/bench_aic_rust_callback.py
index b53bd15721a0..158c65859056 100644
--- a/benchmarks/mocker/bench_aic_rust_callback.py
+++ b/benchmarks/mocker/bench_aic_rust_callback.py
@@ -65,7 +65,7 @@
 MODEL = "Qwen/Qwen3-32B"
 SYSTEM = "h200_sxm"
 BACKEND = "vllm"
-BACKEND_VERSION = "0.19.0"
+BACKEND_VERSION = "current"
 
 # Workloads chosen to stress the predict hot path differently. ``num_workers`` > 1
 # is where the Python GIL serialises predictions, so the Rust win is largest there.
diff --git a/benchmarks/pyproject.toml b/benchmarks/pyproject.toml
index 7b989453d1c0..221802fd115e 100644
--- a/benchmarks/pyproject.toml
+++ b/benchmarks/pyproject.toml
@@ -39,7 +39,7 @@ classifiers = [
 ]
 
 dependencies = [
-    "aisimulate==0.1.0.dev2; python_version >= '3.11' and python_version < '3.14'",
+    "aisimulate==0.12.0.dev1; python_version >= '3.11' and python_version < '3.14'",
     "aiperf==0.12.0",
     "matplotlib",
     "networkx",
diff --git a/components/src/dynamo/mocker/args.py b/components/src/dynamo/mocker/args.py
index ecbc9df49052..6974ccf1b4fa 100644
--- a/components/src/dynamo/mocker/args.py
+++ b/components/src/dynamo/mocker/args.py
@@ -344,8 +344,9 @@ def parse_args(argv: list[str] | None = None) -> argparse.Namespace:
         "--aic-backend-version",
         type=str,
         default=None,
-        help="AIC backend engine version (e.g., '0.19.0' for vLLM, '0.5.10' for SGLang, "
-        "'1.3.0rc10' for TRT-LLM). If not set, uses the default version for the backend.",
+        help="AIC performance-database version: 'current', 'previous', or 'next' "
+        "when available, or a version assigned to one of those slots. "
+        "Defaults to the release database's 'current' slot.",
     )
     parser.add_argument(
         "--aic-tp-size",
diff --git a/components/src/dynamo/planner/tests/integration/test_aic_native_estimator.py b/components/src/dynamo/planner/tests/integration/test_aic_native_estimator.py
index 0237c52dbbda..d00a2d8cc680 100644
--- a/components/src/dynamo/planner/tests/integration/test_aic_native_estimator.py
+++ b/components/src/dynamo/planner/tests/integration/test_aic_native_estimator.py
@@ -40,7 +40,7 @@ def _config() -> PlannerConfig:
             hf_id="Qwen/Qwen3-32B",
             system="h200_sxm",
             backend="vllm",
-            backend_version="0.19.0",
+            backend_version="current",
             prefill_pick=pick,
             decode_pick=pick,
         ),
diff --git a/components/src/dynamo/profiler/tests/integration/test_aic_task_v2_contract.py b/components/src/dynamo/profiler/tests/integration/test_aic_task_v2_contract.py
index 60a82c4faf34..8d9032d5d5e4 100644
--- a/components/src/dynamo/profiler/tests/integration/test_aic_task_v2_contract.py
+++ b/components/src/dynamo/profiler/tests/integration/test_aic_task_v2_contract.py
@@ -31,7 +31,7 @@ def test_rapid_task_v2_build_and_execute_contract() -> None:
         total_gpus=2,
         system="h200_sxm",
         backend="vllm",
-        backend_version="0.19.0",
+        backend_version="current",
         isl=128,
         osl=8,
         ttft=100_000.0,
diff --git a/components/src/dynamo/profiler/tests/unit/test_dgd_generation_aic.py b/components/src/dynamo/profiler/tests/unit/test_dgd_generation_aic.py
index 92452d2305de..5d02988751ff 100644
--- a/components/src/dynamo/profiler/tests/unit/test_dgd_generation_aic.py
+++ b/components/src/dynamo/profiler/tests/unit/test_dgd_generation_aic.py
@@ -313,21 +313,21 @@ def test_injects_all_required_flags(self):
         assert out[out.index("--aic-attention-dp-size") + 1] == "8"
         # trtllm is not a mocker engine_type; leave --engine-type alone.
         assert "--engine-type" not in out
-        assert out[out.index("--aic-backend-version") + 1] == "1.3.0rc10"
+        assert out[out.index("--aic-backend-version") + 1] == "current"
 
     def test_matches_engine_type_for_vllm(self):
         spec = self._spec("vllm")
         out = _inject_mocker_aic_args([], spec, spec.prefill_pick)
         assert out[out.index("--engine-type") + 1] == "vllm"
         assert out[out.index("--aic-backend") + 1] == "vllm"
-        assert out[out.index("--aic-backend-version") + 1] == "0.14.0"
+        assert out[out.index("--aic-backend-version") + 1] == "current"
 
     def test_matches_engine_type_for_sglang(self):
         spec = self._spec("sglang")
         out = _inject_mocker_aic_args([], spec, spec.decode_pick)
         assert out[out.index("--engine-type") + 1] == "sglang"
         assert out[out.index("--aic-backend") + 1] == "sglang"
-        assert out[out.index("--aic-backend-version") + 1] == "0.5.6.post2"
+        assert out[out.index("--aic-backend-version") + 1] == "current"
 
 
 class TestBuildPlannerConfigEmbedsAicSpec:
diff --git a/components/src/dynamo/profiler/tests/unit/test_replay_aic_parity.py b/components/src/dynamo/profiler/tests/unit/test_replay_aic_parity.py
index 9f9722c74df6..775a506b31ee 100644
--- a/components/src/dynamo/profiler/tests/unit/test_replay_aic_parity.py
+++ b/components/src/dynamo/profiler/tests/unit/test_replay_aic_parity.py
@@ -19,8 +19,8 @@
 AIC_PARITY_MODEL = "Qwen/Qwen3-32B"
 AIC_PARITY_SYSTEM = "h200_sxm"
 AIC_PARITY_VERSIONS = {
-    "vllm": "0.19.0",
-    "sglang": "0.5.6.post2",
+    "vllm": "current",
+    "sglang": "current",
 }
 AIC_PARITY_BACKENDS = [
     pytest.param("vllm", id="vllm"),
diff --git a/components/src/dynamo/profiler/utils/dgd_generation.py b/components/src/dynamo/profiler/utils/dgd_generation.py
index 23e8e548453d..f93621369964 100644
--- a/components/src/dynamo/profiler/utils/dgd_generation.py
+++ b/components/src/dynamo/profiler/utils/dgd_generation.py
@@ -23,6 +23,9 @@
 import numpy as np
 import yaml
 
+from dynamo._internal.aic import (
+    DEFAULT_BACKEND_VERSIONS as _MOCKER_AIC_BACKEND_VERSIONS,
+)
 from dynamo.common.utils.paths import get_workspace_dir
 from dynamo.planner.config.aic_interpolation_spec import AICInterpolationSpec
 from dynamo.planner.config.backend_components import MockerComponentName
@@ -54,15 +57,9 @@
     needs_mocker_aic_perf_model,
     needs_profile_data,
 )
-from dynamo.profiler.utils.replay_optimize.constants import AIC_BACKEND_VERSIONS
 
 logger = logging.getLogger(__name__)
 
-_MOCKER_AIC_BACKEND_VERSIONS = {
-    **AIC_BACKEND_VERSIONS,
-    "trtllm": "1.3.0rc10",
-}
-
 
 def _load_latest_database_version() -> Optional[Callable[..., Optional[str]]]:
     try:
diff --git a/components/src/dynamo/profiler/utils/replay_optimize/constants.py b/components/src/dynamo/profiler/utils/replay_optimize/constants.py
index 793b5a68c25d..6d0bd6ac412c 100644
--- a/components/src/dynamo/profiler/utils/replay_optimize/constants.py
+++ b/components/src/dynamo/profiler/utils/replay_optimize/constants.py
@@ -5,9 +5,10 @@
 
 import os
 
+from dynamo._internal.aic import DEFAULT_BACKEND_VERSIONS
+
 AIC_BACKEND_VERSIONS = {
-    "vllm": "0.14.0",
-    "sglang": "0.5.6.post2",
+    backend: DEFAULT_BACKEND_VERSIONS[backend] for backend in ("vllm", "sglang")
 }
 
 DEFAULT_OVERLAP_SCORE_CREDITS = (1.0,)
diff --git a/components/src/dynamo/replay/config.py b/components/src/dynamo/replay/config.py
index 32b1d02f4775..8d9778799b3b 100644
--- a/components/src/dynamo/replay/config.py
+++ b/components/src/dynamo/replay/config.py
@@ -13,6 +13,7 @@
 
 from aisimulate.aic import materialize_aic_num_gpu_blocks
 
+from dynamo._internal.aic import resolve_backend_version
 from dynamo.mocker import MockEngineArgs
 from dynamo.mocker.args import (
     resolve_planner_profile_data as _resolve_mocker_planner_profile_data,
@@ -26,6 +27,13 @@ class PlannerProfileDataResult(Protocol):
 def resolve_aic_num_gpu_blocks(raw: dict[str, Any]) -> None:
     """Materialize AIC KV capacity in-place for SDK compatibility."""
 
+    if (
+        raw.get("aic_backend") is not None
+        or raw.get("aic_attention_dp_size") is not None
+    ):
+        raw["aic_backend_version"] = resolve_backend_version(
+            raw.get("aic_backend") or "vllm", raw.get("aic_backend_version")
+        )
     lowered = materialize_aic_num_gpu_blocks(raw)
     raw.clear()
     raw.update(lowered)
diff --git a/components/src/dynamo/replay/simulation.py b/components/src/dynamo/replay/simulation.py
index 54edaecf3ec2..f10136e89d3e 100644
--- a/components/src/dynamo/replay/simulation.py
+++ b/components/src/dynamo/replay/simulation.py
@@ -17,7 +17,6 @@
 from numbers import Real
 from typing import Any
 
-from aisimulate.aic import materialize_aic_num_gpu_blocks
 from aisimulate.sweeper.provider import JSONValue, RuntimeHookSpec
 from aisimulate.sweeper.replay import (
     HookCapability,
@@ -30,6 +29,7 @@
 from dynamo.llm import AicPerfConfig, KvRouterConfig
 from dynamo.mocker import MockEngineArgs
 from dynamo.replay.api import run_synthetic_trace_replay, run_trace_replay
+from dynamo.replay.config import resolve_aic_num_gpu_blocks
 
 _PLANNER_HOOK = HookCapability(
     provider="dynamo.planner",
@@ -260,7 +260,7 @@ def _engine_args(payload: dict[str, JSONValue] | None) -> MockEngineArgs:
                 "aic_model_path",
             ):
                 lowered.pop(name, None)
-        lowered = materialize_aic_num_gpu_blocks(lowered)
+        resolve_aic_num_gpu_blocks(lowered)
         # Pipeline parallelism is already represented in the public parallel
         # mapping and used for AIC capacity. MockEngineArgs has no PP field.
         lowered.pop("aic_pp_size", None)
diff --git a/components/src/dynamo/replay/tests/test_simulation.py b/components/src/dynamo/replay/tests/test_simulation.py
index a20c93399926..61f7c2ca0778 100644
--- a/components/src/dynamo/replay/tests/test_simulation.py
+++ b/components/src/dynamo/replay/tests/test_simulation.py
@@ -369,7 +369,7 @@ def test_fixed_timing_keeps_aic_identity_out_of_runtime_args(monkeypatch) -> Non
     monkeypatch.setattr(simulation, "MockEngineArgs", _FakeEngineArgs)
     monkeypatch.setattr(
         simulation,
-        "materialize_aic_num_gpu_blocks",
+        "resolve_aic_num_gpu_blocks",
         lambda payload: payload,
     )
     engine_args = simulation.DynamoReplayRunner._engine_args(
diff --git a/components/src/dynamo/replay/tests/test_simulation_integration.py b/components/src/dynamo/replay/tests/test_simulation_integration.py
index 756f5a3c6ede..160bdfcdf5f7 100644
--- a/components/src/dynamo/replay/tests/test_simulation_integration.py
+++ b/components/src/dynamo/replay/tests/test_simulation_integration.py
@@ -235,7 +235,7 @@ def test_sweeper_runs_real_dynamo_replay_in_spawned_workers() -> None:
         include_router=True,
     )
 
-    candidates = Sweeper(
+    result = Sweeper(
         runner_factory=DynamoReplayRunnerFactory(),
         providers={
             "dynamo.planner": create_planner_provider(),
@@ -245,7 +245,9 @@ def test_sweeper_runs_real_dynamo_replay_in_spawned_workers() -> None:
         show_progress=False,
     ).run(config)
 
+    candidates = result.candidates
     assert len(candidates) == 2
+    assert all(candidate.status == "feasible" for candidate in candidates)
     assert all(
         candidate.metrics["output_throughput_tok_s"] > 0.0 for candidate in candidates
     )
diff --git a/container/deps/requirements.aisimulate.txt b/container/deps/requirements.aisimulate.txt
index bf9b45abcf43..7d834b93a593 100644
--- a/container/deps/requirements.aisimulate.txt
+++ b/container/deps/requirements.aisimulate.txt
@@ -2,4 +2,4 @@
 # SPDX-License-Identifier: Apache-2.0
 
 # Published AISimulate wheel staged into planner and framework runtime wheelhouses.
-aisimulate==0.1.0.dev2
+aisimulate==0.12.0.dev1
diff --git a/container/templates/wheel_builder.Dockerfile b/container/templates/wheel_builder.Dockerfile
index c81a67297659..03372c8ca5a4 100644
--- a/container/templates/wheel_builder.Dockerfile
+++ b/container/templates/wheel_builder.Dockerfile
@@ -639,10 +639,22 @@ COPY container/deps/requirements.aisimulate.txt /opt/dynamo/container/deps/requi
 # wheel consumed by ai-dynamo instead of rebuilding it from vendored source.
 # Download only this distribution; runtime images own dependency installation
 # through their requirements files and local wheels.
+# AISimulate targets glibc 2.34+, which all consuming runtime images support.
+# Select that target explicitly: the manylinux_2_28 builder only stages the
+# wheel and must not filter it using the builder's older glibc version.
 RUN --mount=type=cache,id=uv-root-{{ context.dynamo.uv_version }},target=/root/.cache/uv,sharing=shared \
     export UV_CACHE_DIR=/root/.cache/uv && \
     source ${VIRTUAL_ENV}/bin/activate && \
+    case "${TARGETARCH}" in \
+        amd64) AISIMULATE_WHEEL_ARCH=x86_64 ;; \
+        arm64) AISIMULATE_WHEEL_ARCH=aarch64 ;; \
+        *) echo "Unsupported AISimulate target architecture: ${TARGETARCH}" >&2; exit 1 ;; \
+    esac && \
     python -m pip download \
+        --platform "manylinux_2_34_${AISIMULATE_WHEEL_ARCH}" \
+        --python-version "${PYTHON_VERSION}" \
+        --implementation cp \
+        --abi abi3 \
         --only-binary=:all: \
         --no-deps \
         --dest /opt/dynamo/dist \
diff --git a/docs/fern/pages/reference/components/mocker-cli-reference.mdx b/docs/fern/pages/reference/components/mocker-cli-reference.mdx
index c10afd5578cc..b29b9a248884 100644
--- a/docs/fern/pages/reference/components/mocker-cli-reference.mdx
+++ b/docs/fern/pages/reference/components/mocker-cli-reference.mdx
@@ -167,8 +167,9 @@ For the timing-model design, see [Mocker Engine Architecture](../../developer-gu
 </ParamField>
 
 <ParamField path="--aic-backend-version" type="string" default="auto">
-  AIC backend engine version, for example `0.12.0` for vLLM. When unset, uses the default version for
-  the backend.
+  AIC performance-database version. Use `current`, `previous`, or `next` when the slot is available
+  for the selected system and backend, or specify a version assigned to one of those slots.
+  When unset, uses the release database's `current` slot.
 </ParamField>
 
 <ParamField path="--aic-tp-size" type="integer" default="1">
diff --git a/lib/bindings/kvbm/Cargo.lock b/lib/bindings/kvbm/Cargo.lock
index 1f91c48d63e0..bbb8ce02a133 100644
--- a/lib/bindings/kvbm/Cargo.lock
+++ b/lib/bindings/kvbm/Cargo.lock
@@ -34,9 +34,9 @@ dependencies = [
 
 [[package]]
 name = "aisimulate-core"
-version = "0.1.0-dev.2"
+version = "0.12.0-dev.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3aa0422f239fd9b9ca92d4a57fe7069771e158293d266e05a20fde477d15b9af"
+checksum = "d4db6a085073ecb1c2569f796ba44742ed4681f8f4060daaed2456ef4ea9b652"
 dependencies = [
  "anyhow",
  "bincode 1.3.3",
@@ -45,6 +45,7 @@ dependencies = [
  "derive_builder",
  "flate2",
  "indicatif 0.18.4",
+ "log",
  "parquet",
  "pep440_rs",
  "quick_cache",
diff --git a/lib/bindings/python/Cargo.lock b/lib/bindings/python/Cargo.lock
index ed198a024a82..ef4c56d7adf3 100644
--- a/lib/bindings/python/Cargo.lock
+++ b/lib/bindings/python/Cargo.lock
@@ -34,9 +34,9 @@ dependencies = [
 
 [[package]]
 name = "aisimulate-core"
-version = "0.1.0-dev.2"
+version = "0.12.0-dev.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3aa0422f239fd9b9ca92d4a57fe7069771e158293d266e05a20fde477d15b9af"
+checksum = "d4db6a085073ecb1c2569f796ba44742ed4681f8f4060daaed2456ef4ea9b652"
 dependencies = [
  "anyhow",
  "bincode 1.3.3",
@@ -45,6 +45,7 @@ dependencies = [
  "derive_builder",
  "flate2",
  "indicatif 0.18.4",
+ "log",
  "parquet",
  "pep440_rs",
  "pyo3",
diff --git a/lib/bindings/python/Cargo.toml b/lib/bindings/python/Cargo.toml
index c3e27c53fa81..6836f8dab7a7 100644
--- a/lib/bindings/python/Cargo.toml
+++ b/lib/bindings/python/Cargo.toml
@@ -60,7 +60,7 @@ request-trace-s3 = ["dynamo-llm/request-trace-s3"]
 # former AIC crate surface. The `python` feature is required for
 # AicEngineBuilder; this extension supplies the interpreter, so the dependency
 # must not enable `extension-module`.
-aisimulate-core = { version = "=0.1.0-dev.2", optional = true, features = ["python"] }
+aisimulate-core = { version = "=0.12.0-dev.1", optional = true, features = ["python"] }
 dynamo-runtime = { path = "../../runtime" }
 dynamo-mocker = { path = "../../mocker" }
 dynamo-parsers = { version = "8.1.0" }
diff --git a/lib/bindings/python/src/dynamo/_internal/aic.py b/lib/bindings/python/src/dynamo/_internal/aic.py
index 532ed2868540..acb0ea883e20 100644
--- a/lib/bindings/python/src/dynamo/_internal/aic.py
+++ b/lib/bindings/python/src/dynamo/_internal/aic.py
@@ -15,11 +15,11 @@
 # Dynamo's historical default when conditional acceptance rates are omitted.
 _DEFAULT_NEXTN_ACCEPT_RATES = [0.85, 0.3, 0.0, 0.0, 0.0]
 
-# Default backend versions match the AIC-core v0.11.0 perf DB.
+# Resolve defaults through the queryable slots in the pinned AISimulate perf DB.
 DEFAULT_BACKEND_VERSIONS = {
-    "vllm": "0.19.0",
-    "sglang": "0.5.10",
-    "trtllm": "1.3.0rc10",
+    "vllm": "current",
+    "sglang": "current",
+    "trtllm": "current",
 }
 _KV_CAPACITY_BACKENDS = frozenset(DEFAULT_BACKEND_VERSIONS)
 DEFAULT_STATIC_STRIDE = 32
@@ -39,7 +39,7 @@ def _validate_kv_capacity_backend(backend_name: str) -> None:
 
 
 def resolve_backend_version(backend_name: str, backend_version: str | None) -> str:
-    """Return the pinned backend version used for AIC perf lookups."""
+    """Preserve explicit versions; otherwise use the release database current slot."""
     if backend_version is not None:
         return backend_version
     return DEFAULT_BACKEND_VERSIONS.get(backend_name, DEFAULT_BACKEND_VERSIONS["vllm"])
diff --git a/lib/bindings/python/tests/replay/replay_utils.py b/lib/bindings/python/tests/replay/replay_utils.py
index d0cebce559b9..6c4ee67ea672 100644
--- a/lib/bindings/python/tests/replay/replay_utils.py
+++ b/lib/bindings/python/tests/replay/replay_utils.py
@@ -36,17 +36,6 @@
 {"timestamp": 3052, "input_length": 2007, "output_length": 354, "hash_ids": [0, 446, 447, 448]}
 """
 
-AIC_PARITY_MODEL = "Qwen/Qwen3-32B"
-AIC_PARITY_SYSTEM = "h200_sxm"
-AIC_PARITY_VERSIONS = {
-    "vllm": "0.14.0",
-    "sglang": "0.5.6.post2",
-}
-AIC_PARITY_BACKENDS = [
-    pytest.param("vllm", id="vllm"),
-    pytest.param("sglang", id="sglang"),
-]
-
 
 def _require_aisimulate_distribution(*, allow_module_level: bool = False) -> None:
     try:
@@ -302,95 +291,6 @@ def _planner_profile_data_npz_path() -> Path:
     )
 
 
-def _aic_replay_args(backend_name: str):
-    payload = {
-        "block_size": 512,
-        "enable_prefix_caching": True,
-        "enable_chunked_prefill": False,
-        "max_num_seqs": 16,
-        "max_num_batched_tokens": 65536,
-        "num_gpu_blocks": 100000,
-        "speedup_ratio": 1.0,
-        "aic_backend": backend_name,
-        "aic_system": AIC_PARITY_SYSTEM,
-        "aic_backend_version": AIC_PARITY_VERSIONS[backend_name],
-        "aic_tp_size": 1,
-        "aic_model_path": AIC_PARITY_MODEL,
-    }
-    if backend_name == "sglang":
-        payload["engine_type"] = "sglang"
-        payload["sglang"] = {
-            "page_size": 512,
-            "max_prefill_tokens": 65536,
-            "chunked_prefill_size": 65536,
-        }
-    return MockEngineArgs.from_json(json.dumps(payload))
-
-
-def _aic_disagg_replay_args(
-    backend_name: str,
-    *,
-    tp_size: int,
-    is_prefill: bool,
-    max_num_seqs: int,
-    max_num_batched_tokens: int,
-):
-    payload = {
-        "block_size": 512,
-        "enable_prefix_caching": False,
-        "enable_chunked_prefill": False,
-        "max_num_seqs": max_num_seqs,
-        "max_num_batched_tokens": max_num_batched_tokens,
-        "num_gpu_blocks": 50000,
-        "speedup_ratio": 1.0,
-        "aic_backend": backend_name,
-        "aic_system": AIC_PARITY_SYSTEM,
-        "aic_backend_version": AIC_PARITY_VERSIONS[backend_name],
-        "aic_tp_size": tp_size,
-        "aic_model_path": AIC_PARITY_MODEL,
-        "is_prefill": is_prefill,
-        "is_decode": not is_prefill,
-    }
-    if backend_name == "sglang":
-        payload["engine_type"] = "sglang"
-        payload["sglang"] = {
-            "page_size": 512,
-            "max_prefill_tokens": 65536,
-            "chunked_prefill_size": 65536,
-        }
-    return MockEngineArgs.from_json(json.dumps(payload))
-
-
-def _run_aic_static_point(backend_name: str, isl: int, osl: int, batch_size: int):
-    aic_core = pytest.importorskip("aiconfigurator_core")
-
-    database = aic_core.sdk.perf_database.get_database(
-        system=AIC_PARITY_SYSTEM,
-        backend=backend_name,
-        version=AIC_PARITY_VERSIONS[backend_name],
-    )
-    backend = aic_core.sdk.backends.factory.get_backend(backend_name)
-    model = aic_core.sdk.models.get_model(
-        model_path=AIC_PARITY_MODEL,
-        model_config=aic_core.sdk.config.ModelConfig(tp_size=1),
-        backend_name=backend_name,
-    )
-    summary = backend.run_static(
-        model=model,
-        database=database,
-        runtime_config=aic_core.sdk.config.RuntimeConfig(
-            batch_size=batch_size,
-            beam_width=1,
-            isl=isl,
-            osl=osl,
-            prefix=0,
-        ),
-        mode="static",
-        stride=32,
-    )
-    return summary.get_summary_df().to_dict(orient="records")[0]
-
-
 def _planner_profile_data_dir_path() -> Path:
     return (
         Path(__file__).resolve().parents[5]
diff --git a/lib/bindings/python/tests/replay/test_replay_aic_capacity.py b/lib/bindings/python/tests/replay/test_replay_aic_capacity.py
index 244ecb935059..12d050d37b75 100644
--- a/lib/bindings/python/tests/replay/test_replay_aic_capacity.py
+++ b/lib/bindings/python/tests/replay/test_replay_aic_capacity.py
@@ -33,7 +33,7 @@ def _direct_aic_replay_args() -> MockEngineArgs:
             {
                 "engine_type": "trtllm",
                 "aic_backend": "trtllm",
-                "aic_backend_version": "1.3.0rc10",
+                "aic_backend_version": "current",
                 "aic_system": "gb200",
                 "aic_model_path": "meta-llama/Meta-Llama-3.1-8B",
                 "aic_tp_size": 1,
@@ -99,7 +99,7 @@ def fake_estimate_num_gpu_blocks(**kwargs):
             "gpu_memory_utilization": 0.8,
             "mem_fraction_static": None,
             "free_gpu_memory_fraction": None,
-            "backend_version": None,
+            "backend_version": "current",
             "pp_size": 1,
             "moe_tp_size": None,
             "moe_ep_size": None,
diff --git a/lib/bindings/python/tests/replay/test_replay_aic_integration.py b/lib/bindings/python/tests/replay/test_replay_aic_integration.py
index 7941c1d5d095..fdb6ee4b3f0b 100644
--- a/lib/bindings/python/tests/replay/test_replay_aic_integration.py
+++ b/lib/bindings/python/tests/replay/test_replay_aic_integration.py
@@ -1,12 +1,14 @@
 # SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
 # SPDX-License-Identifier: Apache-2.0
 
-"""Offline, network-free coverage of Dynamo replay against AIC 0.11."""
+"""Offline, network-free coverage of Dynamo replay against the pinned AISimulate release."""
 
 from __future__ import annotations
 
 import pytest
 
+from dynamo.replay.config import load_engine_args
+
 pytestmark = [
     pytest.mark.aiconfigurator,
     pytest.mark.gpu_0,
@@ -17,7 +19,7 @@
 
 AIC_MODEL = "Qwen/Qwen3-32B"
 AIC_SYSTEM = "h200_sxm"
-AIC_BACKEND_VERSION = "0.19.0"
+AIC_BACKEND_VERSION = "current"
 
 
 @pytest.fixture(autouse=True)
@@ -61,6 +63,23 @@ def test_real_aic_memory_estimates_gpu_blocks() -> None:
     assert blocks > 0
 
 
+def test_default_aic_capacity_uses_queryable_version() -> None:
+    args = load_engine_args(
+        {
+            "aic_backend": "vllm",
+            "aic_system": AIC_SYSTEM,
+            "aic_model_path": AIC_MODEL,
+            "aic_tp_size": 1,
+            "block_size": 64,
+            "max_num_batched_tokens": 4096,
+            "max_num_seqs": 128,
+        }
+    )
+    assert args is not None
+    assert args.num_gpu_blocks > 0
+    assert args.aic_backend_version == "current"
+
+
 def test_aggregated_replay_uses_native_aic_engine() -> None:
     from aiconfigurator_core.sdk.engine import compile_engine
 
diff --git a/lib/bindings/python/tests/test_aic_capacity.py b/lib/bindings/python/tests/test_aic_capacity.py
index c26bf7973ab2..a4ee02f1fcce 100644
--- a/lib/bindings/python/tests/test_aic_capacity.py
+++ b/lib/bindings/python/tests/test_aic_capacity.py
@@ -264,6 +264,11 @@ def broken_memory_module(name, *args, **kwargs):
     assert exc_info.value is missing_dependency
 
 
+@pytest.mark.parametrize("backend", ["vllm", "sglang", "trtllm"])
+def test_default_version_uses_queryable_slot(backend):
+    assert resolve_backend_version(backend, None) == "current"
+
+
 def test_trtllm_version_resolution():
     assert resolve_backend_version("trtllm", "0.20.0") == "0.20.0"
 
diff --git a/lib/mocker/src/common/protocols.rs b/lib/mocker/src/common/protocols.rs
index bce94d72a27e..9524f05bec9c 100644
--- a/lib/mocker/src/common/protocols.rs
+++ b/lib/mocker/src/common/protocols.rs
@@ -461,7 +461,9 @@ struct MockEngineArgsSerde {
     enable_local_indexer: OptionalConfigValue<bool>,
     bootstrap_port: OptionalConfigValue<u16>,
     handoff_session_timeout_ms: OptionalConfigValue<u64>,
+    #[serde(alias = "kv_transfer_bytes_per_token")]
     kv_bytes_per_token: OptionalConfigValue<usize>,
+    kv_cache_bytes_per_token: OptionalConfigValue<usize>,
     kv_transfer_bandwidth: OptionalConfigValue<f64>,
     kv_transfer_timing_mode: OptionalConfigValue<String>,
     reasoning: OptionalConfigValue<ReasoningConfig>,
@@ -591,8 +593,9 @@ pub struct MockEngineArgs {
     #[builder(default = "None")]
     pub aic_system: Option<String>,
 
-    /// AIC backend engine version (e.g., "0.12.0" for vLLM, "0.5.6.post2" for SGLang).
-    /// If None, uses the default version for the backend.
+    /// AIC performance-database slot ("current", "previous", or "next" when available),
+    /// or a version assigned to one of those slots.
+    /// If None, uses the release database's "current" slot.
     #[serde(skip)]
     #[builder(default = "None")]
     pub aic_backend_version: Option<String>,
@@ -703,11 +706,16 @@ pub struct MockEngineArgs {
     #[validate(range(min = 1))]
     pub handoff_session_timeout_ms: u64,
 
-    /// KV cache bytes per token, auto-computed from model config by Python CLI.
+    /// Bytes transferred per token, auto-computed from model config by Python CLI.
     /// Formula: num_layers * 2 * num_kv_heads * head_dim * dtype_bytes
     #[builder(default = "None")]
     pub kv_bytes_per_token: Option<usize>,
 
+    /// Physical KV-cache bytes occupied by one token, independent of transfer geometry.
+    #[builder(default = "None")]
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub kv_cache_bytes_per_token: Option<usize>,
+
     /// KV cache transfer bandwidth in GB/s for disaggregated serving latency simulation.
     /// Default: 64.0 (inter-node InfiniBand). Set to 0 to disable KV transfer delay.
     /// For intra-node NVLink, typical value is ~450.
@@ -1047,6 +1055,9 @@ impl TryFrom<MockEngineArgsSerde> for MockEngineArgs {
         if let Some(kv_bytes_per_token) = compat.kv_bytes_per_token.into_nullable() {
             builder = builder.kv_bytes_per_token(kv_bytes_per_token);
         }
+        if let Some(kv_cache_bytes_per_token) = compat.kv_cache_bytes_per_token.into_nullable() {
+            builder = builder.kv_cache_bytes_per_token(kv_cache_bytes_per_token);
+        }
         if let Some(kv_transfer_bandwidth) = compat.kv_transfer_bandwidth.into_nullable() {
             builder = builder.kv_transfer_bandwidth(kv_transfer_bandwidth);
         }
diff --git a/lib/mocker/src/engine_adapter.rs b/lib/mocker/src/engine_adapter.rs
index 98ad4c1785b4..a92f415473f9 100644
--- a/lib/mocker/src/engine_adapter.rs
+++ b/lib/mocker/src/engine_adapter.rs
@@ -125,7 +125,8 @@ pub(crate) fn engine_components(
         preemption_mode,
         emit_kv_events,
         emit_kv_token_ids,
-        kv_bytes_per_token: args.kv_bytes_per_token,
+        kv_transfer_bytes_per_token: args.kv_bytes_per_token,
+        kv_cache_bytes_per_token: args.kv_cache_bytes_per_token,
         kv_transfer_bandwidth: args.kv_transfer_bandwidth,
         kv_transfer_timing_mode,
         timing_model,
@@ -159,6 +160,7 @@ pub(crate) fn aggregated_replay_setup(
     let config = ReplayEngineConfig {
         dp_size: components.args.dp_size,
         tensor_parallel_size: replay_tensor_parallel_size(&components.args)?,
+        num_gpu_blocks_is_explicit: None,
         rank: components.rank,
         prefill: None,
         decode: None,
@@ -181,16 +183,19 @@ pub(crate) fn disaggregated_replay_setup(
     let prefill_role = ReplayRoleConfig {
         dp_size: prefill.args.dp_size,
         tensor_parallel_size: replay_tensor_parallel_size(&prefill.args)?,
+        num_gpu_blocks_is_explicit: None,
         rank: prefill.rank,
     };
     let decode_role = ReplayRoleConfig {
         dp_size: decode.args.dp_size,
         tensor_parallel_size: replay_tensor_parallel_size(&decode.args)?,
+        num_gpu_blocks_is_explicit: None,
         rank: decode.rank,
     };
     let config = ReplayEngineConfig {
         dp_size: prefill_role.dp_size,
         tensor_parallel_size: prefill_role.tensor_parallel_size,
+        num_gpu_blocks_is_explicit: prefill_role.num_gpu_blocks_is_explicit,
         rank: prefill_role.rank.clone(),
         prefill: Some(prefill_role),
         decode: Some(decode_role),
@@ -279,18 +284,34 @@ mod tests {
 
     #[test]
     fn vllm_defaults_materialize_once_at_the_shared_boundary() {
-        let args = MockEngineArgs::builder().build().unwrap();
+        let mut args = MockEngineArgs::builder().build().unwrap();
+        args.kv_bytes_per_token = Some(4096);
+        args.kv_cache_bytes_per_token = Some(1024);
         let components = engine_components(args, true, true).unwrap();
 
         assert_eq!(components.args.block_size, 64);
         assert_eq!(components.rank.backend, Backend::Vllm);
         assert_eq!(components.rank.block_size, 64);
+        assert_eq!(components.rank.kv_transfer_bytes_per_token, Some(4096));
+        assert_eq!(components.rank.kv_cache_bytes_per_token, Some(1024));
         assert!(components.rank.emit_kv_events);
         assert!(components.rank.emit_kv_token_ids);
         assert_eq!(components.rank.timing_model, TimingModelConfig::Polynomial);
         assert!(components.timing.is_none());
     }
 
+    #[test]
+    fn replay_json_keeps_transfer_and_cache_geometry_independent() {
+        let args: MockEngineArgs = serde_json::from_value(serde_json::json!({
+            "kv_transfer_bytes_per_token": 4096,
+            "kv_cache_bytes_per_token": 1024,
+        }))
+        .unwrap();
+        let components = engine_components(args, false, false).unwrap();
+        assert_eq!(components.rank.kv_transfer_bytes_per_token, Some(4096));
+        assert_eq!(components.rank.kv_cache_bytes_per_token, Some(1024));
+    }
+
     #[test]
     fn backend_specific_fields_match_the_engine_contract() {
         let mut sglang = MockEngineArgs::builder().build().unwrap();
diff --git a/lib/mocker/src/replay/offline/extensions/kv_router/mod.rs b/lib/mocker/src/replay/offline/extensions/kv_router/mod.rs
index 5698062371c7..4ea2d4de1258 100644
--- a/lib/mocker/src/replay/offline/extensions/kv_router/mod.rs
+++ b/lib/mocker/src/replay/offline/extensions/kv_router/mod.rs
@@ -154,6 +154,7 @@ pub(crate) struct WorkerAdmission {
     uuid: Uuid,
     worker_idx: usize,
     overlap_blocks: u32,
+    best_available_overlap_blocks: u32,
     isl_blocks: u32,
 }
 
@@ -169,6 +170,7 @@ pub(crate) struct RouterEffects {
 struct AdmitOutcome {
     worker_idx: usize,
     overlap_blocks: u32,
+    best_available_overlap_blocks: u32,
     isl_blocks: u32,
 }
 
@@ -368,10 +370,12 @@ impl KvRouterPlacement {
         Placement {
             request_id: admission.uuid,
             scheduler_id: admission.worker_idx,
+            placement_replica_id: None,
             reported_overlap_tokens: admission.overlap_blocks as usize
                 * self.router.block_size as usize,
             cache_sample: Some(PlacementCacheSample {
                 overlap_blocks: admission.overlap_blocks,
+                best_available_overlap_blocks: admission.best_available_overlap_blocks,
                 isl_blocks: admission.isl_blocks,
             }),
         }
@@ -662,6 +666,7 @@ impl OfflineReplayRouter {
                 uuid,
                 worker_idx: outcome.worker_idx,
                 overlap_blocks: outcome.overlap_blocks,
+                best_available_overlap_blocks: outcome.best_available_overlap_blocks,
                 isl_blocks: outcome.isl_blocks,
             }],
         })
@@ -927,6 +932,19 @@ impl OfflineReplayRouter {
             .project_worker_loads(request.token_seq.as_deref(), decay_now);
         let scheduling_request = request.scheduling_request(self.block_size as usize, worker_loads);
         let eligibility = scheduling_request.eligibility();
+        let best_available_overlap_blocks = request
+            .overlaps
+            .scores
+            .iter()
+            .filter(|(worker, _)| {
+                self.workers_with_configs
+                    .get(&worker.worker_id)
+                    .is_some_and(|config| eligibility.allows_worker(worker.worker_id, config))
+                    && worker.dp_rank < self.dp_size
+            })
+            .map(|(_, overlap)| *overlap)
+            .max()
+            .unwrap_or(0);
         let selection = self
             .selector
             .select_worker(WorkerSelectionInput::configured(
@@ -972,6 +990,7 @@ impl OfflineReplayRouter {
         Ok(AdmitOutcome {
             worker_idx,
             overlap_blocks,
+            best_available_overlap_blocks,
             isl_blocks,
         })
     }
@@ -993,6 +1012,7 @@ impl OfflineReplayRouter {
                 uuid,
                 worker_idx: outcome.worker_idx,
                 overlap_blocks: outcome.overlap_blocks,
+                best_available_overlap_blocks: outcome.best_available_overlap_blocks,
                 isl_blocks: outcome.isl_blocks,
             });
         }
@@ -1419,6 +1439,7 @@ mod tests {
                 uuid: Uuid::from_u128(1),
                 worker_idx: 1,
                 overlap_blocks: 1,
+                best_available_overlap_blocks: 1,
                 isl_blocks: 1,
             }]
         );
@@ -1804,11 +1825,37 @@ policy_classes:
                 uuid: Uuid::from_u128(1),
                 worker_idx: 3,
                 overlap_blocks: 0,
+                best_available_overlap_blocks: 0,
                 isl_blocks: 1,
             }]
         );
     }
 
+    #[test]
+    fn cache_telemetry_excludes_removed_workers() {
+        let mut router =
+            OfflineReplayRouter::new(&replay_args(), Some(router_config()), None, 2).unwrap();
+        let target = request(1, 7);
+        let hashes = ReplayRequestHashes::from_tokens(&target.tokens, router.block_size);
+        router
+            .on_kv_events(vec![store_event(
+                1,
+                1,
+                hashes.local_block_hashes[0],
+                StorageTier::Device,
+            )])
+            .unwrap();
+        router.remove_worker(1).unwrap();
+
+        let effects = router
+            .on_request_arrival(&target, Some(hashes), 0.0)
+            .unwrap();
+        assert_eq!(effects.admissions.len(), 1);
+        assert_eq!(effects.admissions[0].worker_idx, 0);
+        assert_eq!(effects.admissions[0].overlap_blocks, 0);
+        assert_eq!(effects.admissions[0].best_available_overlap_blocks, 0);
+    }
+
     #[test]
     fn finalized_worker_removal_drops_retained_topology_and_cache_state() {
         let mut router =
@@ -1901,6 +1948,7 @@ policy_classes:
                 uuid: Uuid::from_u128(2),
                 worker_idx: 1,
                 overlap_blocks: 0,
+                best_available_overlap_blocks: 0,
                 isl_blocks: 1,
             }]
         );
diff --git a/pyproject.toml b/pyproject.toml
index 31c19289007c..50c0f3d73ccc 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -14,7 +14,7 @@ license-files = ["LICENSE"]
 requires-python = ">=3.10"
 dependencies = [
     "ai-dynamo-runtime==1.5.0",
-    "aisimulate==0.1.0.dev2; python_version >= '3.11' and python_version < '3.14'",
+    "aisimulate==0.12.0.dev1; python_version >= '3.11' and python_version < '3.14'",
     "aiohttp>=3.14.3,<4.0",
     "transformers>=4.56.0",
     "kubernetes>=32.0.1,<33.0.0",
@@ -240,6 +240,9 @@ xfail_strict = true
 log_cli_level = "INFO"
 filterwarnings = [
     "error",
+    # DGDR still uses AISimulate 0.12's supported legacy sweep API until its
+    # migration to the new Sweeper API is implemented.
+    'ignore:`aiconfigurator\.sdk\.sweep\.sweep_(agg|disagg)\(\)` is deprecated:DeprecationWarning',
     # chardet 6.x is pulled in transitively but requests 2.32.x only supports chardet <6;
     # charset_normalizer is used at runtime so the warning is benign.
     "ignore:.*doesn't match a supported version",
diff --git a/tests/dependencies/test_aiconfigurator_consistency.py b/tests/dependencies/test_aiconfigurator_consistency.py
index 332b3c0e59a7..80c27c7c2358 100644
--- a/tests/dependencies/test_aiconfigurator_consistency.py
+++ b/tests/dependencies/test_aiconfigurator_consistency.py
@@ -93,7 +93,7 @@ def test_no_manifest_installs_retired_aic_distributions() -> None:
         assert all(package["name"] != "aiconfigurator-core" for package in packages)
     assert features["aic-forward-pass"] == ["dep:aisimulate-core"]
     assert dependencies["aisimulate-core"] == {
-        "version": "=0.1.0-dev.2",
+        "version": "=0.12.0-dev.1",
         "optional": True,
         "features": ["python"],
     }
diff --git a/tests/frontend/router_aic_image_smoke.py b/tests/frontend/router_aic_image_smoke.py
index 8a85f48b6029..4a7769fe68f2 100644
--- a/tests/frontend/router_aic_image_smoke.py
+++ b/tests/frontend/router_aic_image_smoke.py
@@ -31,7 +31,7 @@ def main() -> None:
         "--aic-model-path",
         "Qwen/Qwen3-32B",
         "--aic-backend-version",
-        "0.19.0",
+        "current",
     ]
 
     from aiconfigurator_core.sdk.engine import EngineHandle
diff --git a/tests/router/test_router_e2e_with_mockers.py b/tests/router/test_router_e2e_with_mockers.py
index a03f083e4623..113a175046ab 100644
--- a/tests/router/test_router_e2e_with_mockers.py
+++ b/tests/router/test_router_e2e_with_mockers.py
@@ -97,7 +97,7 @@
 ROUTER_AIC_CONFIG = {
     "aic_backend": "vllm",
     "aic_system": "h200_sxm",
-    "aic_backend_version": "0.14.0",
+    "aic_backend_version": "current",
     "aic_tp_size": 1,
     "aic_model_path": "Qwen/Qwen3-32B",
 }
diff --git a/tests/wheels/smoke_install.py b/tests/wheels/smoke_install.py
index 6f82181f8624..d645cd1c8194 100644
--- a/tests/wheels/smoke_install.py
+++ b/tests/wheels/smoke_install.py
@@ -406,7 +406,7 @@ def run_aic_core_import_smoke(venv_python: Path) -> None:
 assert (package_root / "model_configs/Qwen--Qwen3-32B_config.json").is_file()
 assert (package_root / "systems/h200_sxm.yaml").is_file()
 parquet_files = list(
-    (package_root / "systems/data/h200_sxm").glob("*/vllm/0.19.0/*.parquet")
+    (package_root / "systems/data/h200_sxm").glob("*/vllm/0.24.0/*.parquet")
 )
 assert parquet_files
 for path in parquet_files:
@@ -419,7 +419,7 @@ def run_aic_core_import_smoke(venv_python: Path) -> None:
         "model_name": "Qwen/Qwen3-32B",
         "system_name": "h200_sxm",
         "backend": "vllm",
-        "backend_version": "0.19.0",
+        "backend_version": "0.24.0",
         "kv_block_size": None,
         "tp_size": 1,
         "pp_size": 1,