diff --git a/litellm/llms/anthropic/prompt_cache_prediction.py b/litellm/llms/anthropic/prompt_cache_prediction.py
index e69a02bd93a..a0ce5bf0360 100644
--- a/litellm/llms/anthropic/prompt_cache_prediction.py
+++ b/litellm/llms/anthropic/prompt_cache_prediction.py
@@ -50,6 +50,7 @@ _DEPLOYMENT_OPTIONS: Final = frozenset(
         "max_retries",
         "num_retries",
         "max_parallel_requests",
+        "max_parallel_requests_queue_size",
         "input_cost_per_token",
         "output_cost_per_token",
         "cache_read_input_token_cost",
diff --git a/tests/test_litellm/llms/anthropic/test_anthropic_prompt_cache_prediction.py b/tests/test_litellm/llms/anthropic/test_anthropic_prompt_cache_prediction.py
index 2b36866a1a0..c13217a0d46 100644
--- a/tests/test_litellm/llms/anthropic/test_anthropic_prompt_cache_prediction.py
+++ b/tests/test_litellm/llms/anthropic/test_anthropic_prompt_cache_prediction.py
@@ -181,6 +181,15 @@ async def test_environment_credential_matches_native_count_and_observed_scope(
     assert observed.scope == cache_scope(_CALLER, _DEPLOYMENT, target.api_key, target.model)
 
 
+def test_deployment_concurrency_knobs_keep_native_prediction_supported() -> None:
+    target: Final = resolve_prediction_target(LiteLLM_Params(
+        model=f"anthropic/{_MODEL}", api_key=_KEY, api_base="https://api.anthropic.com",
+        max_parallel_requests=1, max_parallel_requests_queue_size=0,
+    ))
+    assert isinstance(target, NativePredictionTarget)
+    assert (target.model, target.api_key) == (_MODEL, _KEY)
+
+
 @pytest.mark.parametrize("inline_key", [None, _KEY])
 @pytest.mark.asyncio
 async def test_named_credential_is_explicitly_unsupported_before_count(