diff --git a/requirements/kv_connectors.txt b/requirements/kv_connectors.txt
index f60a01a55d07c..083230c171096 100644
--- a/requirements/kv_connectors.txt
+++ b/requirements/kv_connectors.txt
@@ -1,2 +1,2 @@
-lmcache >= 0.3.10.post1
+lmcache
 nixl >= 0.7.1 # Required for disaggregated prefill
diff --git a/vllm/distributed/kv_transfer/kv_connector/v1/lmcache_integration/vllm_v1_adapter.py b/vllm/distributed/kv_transfer/kv_connector/v1/lmcache_integration/vllm_v1_adapter.py
index cdc2969a7735e..09af128f3ed74 100644
--- a/vllm/distributed/kv_transfer/kv_connector/v1/lmcache_integration/vllm_v1_adapter.py
+++ b/vllm/distributed/kv_transfer/kv_connector/v1/lmcache_integration/vllm_v1_adapter.py
@@ -27,7 +27,14 @@ from lmcache.v1.lookup_client.lmcache_async_lookup_client import (
     LMCacheAsyncLookupServer,
 )
 from lmcache.v1.offload_server.zmq_server import ZMQOffloadServer
-from lmcache.v1.plugin.runtime_plugin_launcher import RuntimePluginLauncher
+
+try:
+    from lmcache.v1.plugin.runtime_plugin_launcher import RuntimePluginLauncher
+except ImportError:
+    # Backwards compatibility for lmcache <= 0.3.10-post1
+    from lmcache.v1.plugin.plugin_launcher import (
+        PluginLauncher as RuntimePluginLauncher,
+    )
 
 from vllm.attention.backends.abstract import AttentionMetadata
 from vllm.config import VllmConfig