diff --git a/Dockerfile.ci b/Dockerfile.ci
index 10ea68f2c247..d51f79e1b3af 100644
--- a/Dockerfile.ci
+++ b/Dockerfile.ci
@@ -53,7 +53,7 @@ RUN pip install nemo_run@git+https://github.com/NVIDIA/NeMo-Run.git@${NEMO_RUN_T
 # Install NeMo requirements
 ARG TE_TAG=7d576ed25266a17a7b651f2c12e8498f67e0baea
 ARG MODELOPT_VERSION=0.19.0
-ARG MCORE_TAG=213c8a23fa9fe95d19eff0932a1e6e71767f0962
+ARG MCORE_TAG=441cb9250101cf2cc406f0439b802f34f923f251
 
 ARG APEX_TAG=810ffae374a2b9cb4b5c5e28eaeca7d7998fca0c
 RUN \
diff --git a/nemo/collections/nlp/models/language_modeling/megatron/bert/bert_model.py b/nemo/collections/nlp/models/language_modeling/megatron/bert/bert_model.py
index 67a4802d83f6..0d75ab7cc706 100644
--- a/nemo/collections/nlp/models/language_modeling/megatron/bert/bert_model.py
+++ b/nemo/collections/nlp/models/language_modeling/megatron/bert/bert_model.py
@@ -206,6 +206,8 @@ def forward(
         context=None,
         context_mask=None,
         rotary_pos_emb=None,
+        rotary_pos_cos=None,
+        rotary_pos_sin=None,
         inference_params=None,
         packed_seq_params=None,
     ):
diff --git a/nemo/collections/nlp/models/language_modeling/megatron/falcon/falcon_decoder_layer.py b/nemo/collections/nlp/models/language_modeling/megatron/falcon/falcon_decoder_layer.py
index 1783d5f5f3fd..131f154d6709 100644
--- a/nemo/collections/nlp/models/language_modeling/megatron/falcon/falcon_decoder_layer.py
+++ b/nemo/collections/nlp/models/language_modeling/megatron/falcon/falcon_decoder_layer.py
@@ -55,7 +55,7 @@ class FalconTransformerLayer(TransformerLayer):
 
     Transformer layer takes input with size [s, b, h] and returns an
     output of the same size.
-        
+
     """
 
     def __init__(
@@ -106,6 +106,8 @@ def forward(
         context=None,
         context_mask=None,
         rotary_pos_emb=None,
+        rotary_pos_cos=None,
+        rotary_pos_sin=None,
         inference_params=None,
         packed_seq_params=None,
     ):
diff --git a/nemo/collections/nlp/models/language_modeling/megatron/gpt_full_te_layer_autocast_spec.py b/nemo/collections/nlp/models/language_modeling/megatron/gpt_full_te_layer_autocast_spec.py
index e9fb1833fc08..d1945139dee9 100644
--- a/nemo/collections/nlp/models/language_modeling/megatron/gpt_full_te_layer_autocast_spec.py
+++ b/nemo/collections/nlp/models/language_modeling/megatron/gpt_full_te_layer_autocast_spec.py
@@ -250,6 +250,8 @@ def forward(
         context=None,
         context_mask=None,
         rotary_pos_emb=None,
+        rotary_pos_cos=None,
+        rotary_pos_sin=None,
         inference_params=None,
         packed_seq_params=None,  # TODO: handle this
     ):
diff --git a/nemo/collections/nlp/models/language_modeling/megatron/griffin/griffin_model.py b/nemo/collections/nlp/models/language_modeling/megatron/griffin/griffin_model.py
index 7a327a3a35cb..e0e3a2339ca1 100755
--- a/nemo/collections/nlp/models/language_modeling/megatron/griffin/griffin_model.py
+++ b/nemo/collections/nlp/models/language_modeling/megatron/griffin/griffin_model.py
@@ -160,7 +160,9 @@ def forward(
         rotary_pos_emb = None
         self.decoder.input_tensor = None
         if self.position_embedding_type == 'rope':
-            rotary_seq_len = self.rotary_pos_emb.get_rotary_seq_len(None, self.decoder, hidden_states, self.config)
+            rotary_seq_len = self.rotary_pos_emb.get_rotary_seq_len(
+                None, self.decoder, hidden_states, self.config, None
+            )
             rotary_pos_emb = self.rotary_pos_emb(rotary_seq_len)
 
         hidden_states = self.decoder(hidden_states, attention_mask=attention_mask, rotary_pos_emb=rotary_pos_emb)
diff --git a/nemo/collections/nlp/modules/common/megatron/adapters/mcore_mixins.py b/nemo/collections/nlp/modules/common/megatron/adapters/mcore_mixins.py
index 5128b4ca6b16..da9c98fd94ea 100644
--- a/nemo/collections/nlp/modules/common/megatron/adapters/mcore_mixins.py
+++ b/nemo/collections/nlp/modules/common/megatron/adapters/mcore_mixins.py
@@ -80,11 +80,21 @@ def forward(
         context: Tensor = None,
         context_mask: Tensor = None,
         rotary_pos_emb: Tensor = None,
+        rotary_pos_cos: Tensor = None,
+        rotary_pos_sin: Tensor = None,
         inference_params: InferenceParams = None,
         packed_seq_params: PackedSeqParams = None,
     ):
         hidden_states = super().forward(
-            hidden_states, attention_mask, context, context_mask, rotary_pos_emb, inference_params, packed_seq_params
+            hidden_states,
+            attention_mask,
+            context,
+            context_mask,
+            rotary_pos_emb,
+            rotary_pos_cos,
+            rotary_pos_sin,
+            inference_params,
+            packed_seq_params,
         )
 
         mlp_head_adapter = self.get_adapter_module(AdapterName.MLP_HEAD_ADAPTER)
@@ -220,6 +230,8 @@ def forward(
         inference_params=None,
         rotary_pos_emb=None,
         packed_seq_params=None,
+        rotary_pos_cos=None,
+        rotary_pos_sin=None,
     ):
         # hidden_states: [sq, b, h]
 
@@ -237,8 +249,8 @@ def forward(
         # ===================================================
         # Adjust key, value, and rotary_pos_emb for inference
         # ===================================================
-        key, value, rotary_pos_emb, attn_mask_type = self._adjust_key_value_for_inference(
-            inference_params, key, value, rotary_pos_emb
+        query, key, value, rotary_pos_emb, attn_mask_type = self._adjust_key_value_for_inference(
+            inference_params, query, key, value, rotary_pos_emb
         )
 
         if packed_seq_params is not None: