diff --git a/include/onnxruntime/core/framework/op_kernel.h b/include/onnxruntime/core/framework/op_kernel.h
index 8ec94c67cc0a4..42e8e9c5e3cbe 100644
--- a/include/onnxruntime/core/framework/op_kernel.h
+++ b/include/onnxruntime/core/framework/op_kernel.h
@@ -105,9 +105,9 @@ class OpKernel {
     return Status::OK();
   }
 
-  // Note: New implementations should override OpKernel::UseSharedPrePackedBuffers_V2 instead.
   // Override this function to use provided pre-packed weight.
   // Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+  //                                 gsl::span<const size_t> prepacked_buffer_sizes,
   //                                 int input_idx,
   //                                 /*out*/ bool& used_shared_buffers) {
   //     used_shared_buffers = true;
@@ -121,37 +121,18 @@ class OpKernel {
   //                            and must use the same order for retrieval in UseSharedPrePackedBuffers(). Though each element
   //                           of this vector is a BufferUniquePtr, the deleter of the BufferUniquePtr is NULL. So actually they
   //                           are raw pointers.
+  // @param prepacked_buffer_sizes: The sizes (in bytes) of each buffer in prepacked_buffers.
   // @param input_idx: The input index of the tensor in this kernel
   // @param used_shared_buffers: Boolean flag set by the kernel implementation indicating
   // that the provided weight has been used by the kernel.
   virtual Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& /*prepacked_buffers*/,
+                                           gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                            int /*input_idx*/,
                                            /*out*/ bool& used_shared_buffers) {
     used_shared_buffers = false;
     return Status::OK();
   }
 
-  /// <summary>
-  /// Version 2 of OpKernel::UseSharedPrePackedBuffers() that additionally accepts the buffer sizes as a parameter.
-  /// The default implementation of this function just calls directly to OpKernel::UseSharedPrePackedBuffers()
-  /// to avoid the need to update all existing kernel-based provider-bridge EPs.
-  ///
-  /// TODO: Consolidate UseSharedPrePackedBuffers and UseSharedPrePackedBuffers_V2 into a single function,
-  /// which will require updating kernel-based provider-bridge EPs (cpu, cuda, webgpu).
-  ///
-  /// </summary>
-  /// <param name="prepacked_buffers"></param>
-  /// <param name="prepacked_buffer_sizes"></param>
-  /// <param name="input_idx"></param>
-  /// <param name="used_shared_buffers"></param>
-  /// <returns></returns>
-  virtual Status UseSharedPrePackedBuffers_V2(std::vector<BufferUniquePtr>& prepacked_buffers,
-                                              gsl::span<const size_t> /*prepacked_buffer_sizes*/,
-                                              int input_idx,
-                                              /*out*/ bool& used_shared_buffers) {
-    return UseSharedPrePackedBuffers(prepacked_buffers, input_idx, used_shared_buffers);
-  }
-
   const OrtDevice GetDevice(OrtMemType mem_type) const;
   const OpKernelInfo& Info() const {
     return *op_kernel_info_;
diff --git a/onnxruntime/contrib_ops/cpu/bert/attention.cc b/onnxruntime/contrib_ops/cpu/bert/attention.cc
index 7268b32623b95..e1981fb5c2442 100644
--- a/onnxruntime/contrib_ops/cpu/bert/attention.cc
+++ b/onnxruntime/contrib_ops/cpu/bert/attention.cc
@@ -34,6 +34,7 @@ class Attention : public OpKernel, public AttentionCPUBase {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
@@ -176,6 +177,7 @@ Status Attention<T>::PrePack(const Tensor& weights, int input_idx, AllocatorPtr
 
 template <typename T>
 Status Attention<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                               gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                int input_idx,
                                                /*out*/ bool& used_shared_buffers) {
   if (1 != input_idx) {
diff --git a/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.cc b/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.cc
index ca2c3ab001da6..a674d05b6daae 100644
--- a/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.cc
+++ b/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.cc
@@ -578,10 +578,10 @@ Status QMoECPU<T>::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr all
 }
 
 template <typename T>
-Status QMoECPU<T>::UseSharedPrePackedBuffers_V2(std::vector<BufferUniquePtr>& prepacked_buffers,
-                                                gsl::span<const size_t> /*prepacked_buffer_sizes*/,
-                                                int input_idx,
-                                                /*out*/ bool& used_shared_buffers) {
+Status QMoECPU<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                             gsl::span<const size_t> /*prepacked_buffer_sizes*/,
+                                             int input_idx,
+                                             /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
 
   if (expert_weight_bits_ != 4) {
@@ -1577,11 +1577,11 @@ template QMoECPU<float>::QMoECPU(const OpKernelInfo& op_kernel_info);
 
 template Status QMoECPU<float>::Compute(OpKernelContext* context) const;
 template Status QMoECPU<float>::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr alloc, bool& is_packed, PrePackedWeights* prepacked_weights);
-template Status QMoECPU<float>::UseSharedPrePackedBuffers_V2(std::vector<BufferUniquePtr>& prepacked_buffers, gsl::span<const size_t> prepacked_buffer_sizes, int input_idx, bool& used_shared_buffers);
+template Status QMoECPU<float>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers, gsl::span<const size_t> prepacked_buffer_sizes, int input_idx, bool& used_shared_buffers);
 template QMoECPU<MLFloat16>::QMoECPU(const OpKernelInfo& op_kernel_info);
 template Status QMoECPU<MLFloat16>::Compute(OpKernelContext* context) const;
 template Status QMoECPU<MLFloat16>::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr alloc, bool& is_packed, PrePackedWeights* prepacked_weights);
-template Status QMoECPU<MLFloat16>::UseSharedPrePackedBuffers_V2(std::vector<BufferUniquePtr>& prepacked_buffers, gsl::span<const size_t> prepacked_buffer_sizes, int input_idx, bool& used_shared_buffers);
+template Status QMoECPU<MLFloat16>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers, gsl::span<const size_t> prepacked_buffer_sizes, int input_idx, bool& used_shared_buffers);
 
 // Kernel Registration
 ONNX_OPERATOR_TYPED_KERNEL_EX(
diff --git a/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.h b/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.h
index f678a27190c90..c5e6904ae48c2 100644
--- a/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.h
+++ b/onnxruntime/contrib_ops/cpu/moe/moe_quantization_cpu.h
@@ -32,10 +32,10 @@ class QMoECPU final : public OpKernel, public MoEBaseCPU {
                  /*out*/ bool& is_packed,
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
-  Status UseSharedPrePackedBuffers_V2(std::vector<BufferUniquePtr>& prepacked_buffers,
-                                      gsl::span<const size_t> prepacked_buffer_sizes,
-                                      int input_idx,
-                                      /*out*/ bool& used_shared_buffers) override;
+  Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> prepacked_buffer_sizes,
+                                   int input_idx,
+                                   /*out*/ bool& used_shared_buffers) override;
 
   void ApplyActivationVectorized(float* data, int64_t size) const;
 
diff --git a/onnxruntime/contrib_ops/cpu/quantization/attention_quant.cc b/onnxruntime/contrib_ops/cpu/quantization/attention_quant.cc
index b30fa1e5e618a..931677582d469 100644
--- a/onnxruntime/contrib_ops/cpu/quantization/attention_quant.cc
+++ b/onnxruntime/contrib_ops/cpu/quantization/attention_quant.cc
@@ -28,6 +28,7 @@ class QAttention : public OpKernel, public AttentionCPUBase {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
@@ -117,6 +118,7 @@ Status QAttention<T>::PrePack(const Tensor& weights, int input_idx, AllocatorPtr
 
 template <typename T>
 Status QAttention<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                 int input_idx,
                                                 /*out*/ bool& used_shared_buffers) {
   if (1 != input_idx) {
diff --git a/onnxruntime/contrib_ops/cpu/quantization/dynamic_quantize_lstm.cc b/onnxruntime/contrib_ops/cpu/quantization/dynamic_quantize_lstm.cc
index f55e66f9c5d81..2094af78f40b7 100644
--- a/onnxruntime/contrib_ops/cpu/quantization/dynamic_quantize_lstm.cc
+++ b/onnxruntime/contrib_ops/cpu/quantization/dynamic_quantize_lstm.cc
@@ -17,6 +17,7 @@ class DynamicQuantizeLSTM : public OpKernel, public LSTMBase {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
@@ -117,6 +118,7 @@ Status DynamicQuantizeLSTM::PrePack(const Tensor& tensor, int input_idx, Allocat
 }
 
 Status DynamicQuantizeLSTM::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                      gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                       int input_idx,
                                                       /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
diff --git a/onnxruntime/contrib_ops/cpu/quantization/matmul_nbits.cc b/onnxruntime/contrib_ops/cpu/quantization/matmul_nbits.cc
index d2996b122c5f7..3da0ee19d4cde 100644
--- a/onnxruntime/contrib_ops/cpu/quantization/matmul_nbits.cc
+++ b/onnxruntime/contrib_ops/cpu/quantization/matmul_nbits.cc
@@ -135,7 +135,9 @@ class MatMulNBits final : public OpKernel {
                  /*out*/ bool& is_packed,
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
-  Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers, int input_idx,
+  Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
+                                   int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
  private:
@@ -557,7 +559,9 @@ Status MatMulNBits<MLFloat16>::PrePack(const Tensor& tensor, int input_idx, /*ou
 #endif  // end !MLAS_F16VEC_INTRINSICS_SUPPORTED || !MLAS_TARGET_ARM64
 
 template <typename T1>
-Status MatMulNBits<T1>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers, int input_idx,
+Status MatMulNBits<T1>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                  gsl::span<const size_t> /*prepacked_buffer_sizes*/,
+                                                  int input_idx,
                                                   /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
 
diff --git a/onnxruntime/core/framework/session_state.cc b/onnxruntime/core/framework/session_state.cc
index 5c33a621cf514..84521af2d8532 100644
--- a/onnxruntime/core/framework/session_state.cc
+++ b/onnxruntime/core/framework/session_state.cc
@@ -436,8 +436,8 @@ static Status KernelUseSharedPrePackedBuffers(OpKernel& kernel, int input_idx,
   }
 
   bool used_shared_buffers = false;
-  ORT_RETURN_IF_ERROR(kernel.UseSharedPrePackedBuffers_V2(shared_prepacked_buffers, shared_prepacked_buffer_sizes,
-                                                          input_idx, used_shared_buffers));
+  ORT_RETURN_IF_ERROR(kernel.UseSharedPrePackedBuffers(shared_prepacked_buffers, shared_prepacked_buffer_sizes,
+                                                       input_idx, used_shared_buffers));
 
   // BUG CHECK: Ensure that the kernel used the provided shared buffers
   // Mostly a debug check to ensure that the kernel has an overridden implementation of the
diff --git a/onnxruntime/core/providers/acl/math/matmul.cc b/onnxruntime/core/providers/acl/math/matmul.cc
index 468b394471c13..029a9ebe2768a 100644
--- a/onnxruntime/core/providers/acl/math/matmul.cc
+++ b/onnxruntime/core/providers/acl/math/matmul.cc
@@ -269,6 +269,7 @@ Status MatMul::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr alloc,
 }
 
 Status MatMul::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                         gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                          int input_idx, /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
   if (input_idx != 1) {
diff --git a/onnxruntime/core/providers/acl/math/matmul.h b/onnxruntime/core/providers/acl/math/matmul.h
index b137e33833de9..783e15585ebf5 100644
--- a/onnxruntime/core/providers/acl/math/matmul.h
+++ b/onnxruntime/core/providers/acl/math/matmul.h
@@ -34,6 +34,7 @@ class MatMul : public OpKernel {
                  bool& is_packed, PrePackedWeights*) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>&,
+                                   gsl::span<const size_t>,
                                    int, bool&) override;
 
   Status Compute(OpKernelContext* context) const override;
diff --git a/onnxruntime/core/providers/acl/nn/conv.cc b/onnxruntime/core/providers/acl/nn/conv.cc
index a62158f1c26ee..5cc10f7cfd2a8 100644
--- a/onnxruntime/core/providers/acl/nn/conv.cc
+++ b/onnxruntime/core/providers/acl/nn/conv.cc
@@ -370,6 +370,7 @@ Status Conv::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr alloc,
 }
 
 Status Conv::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                       gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                        int input_idx, /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
   if (isQuantized ? (input_idx != 3) : (input_idx != 1)) {
diff --git a/onnxruntime/core/providers/acl/nn/conv.h b/onnxruntime/core/providers/acl/nn/conv.h
index b05ba5363542f..7af086a410857 100644
--- a/onnxruntime/core/providers/acl/nn/conv.h
+++ b/onnxruntime/core/providers/acl/nn/conv.h
@@ -36,6 +36,7 @@ class Conv : public onnxruntime::OpKernel {
                  bool& is_packed, PrePackedWeights*) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>&,
+                                   gsl::span<const size_t>,
                                    int, bool&) override;
 
   Status Compute(OpKernelContext* context) const override;
diff --git a/onnxruntime/core/providers/cpu/fp16/fp16_conv.cc b/onnxruntime/core/providers/cpu/fp16/fp16_conv.cc
index 790b1543bbd74..08dbc46213f65 100644
--- a/onnxruntime/core/providers/cpu/fp16/fp16_conv.cc
+++ b/onnxruntime/core/providers/cpu/fp16/fp16_conv.cc
@@ -54,6 +54,7 @@ class FusedConvFp16 final : public OpKernel {
                  /*out*/ bool& is_packed, /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
@@ -211,6 +212,7 @@ Status FusedConvFp16::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr
 }
 
 Status FusedConvFp16::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                 int input_idx,
                                                 /*out*/ bool& used_shared_buffers) {
   if (input_idx != 1) {
diff --git a/onnxruntime/core/providers/cpu/math/gemm.cc b/onnxruntime/core/providers/cpu/math/gemm.cc
index ac931c76ee3ae..c0da9aec1e1b1 100644
--- a/onnxruntime/core/providers/cpu/math/gemm.cc
+++ b/onnxruntime/core/providers/cpu/math/gemm.cc
@@ -296,6 +296,7 @@ Status Gemm<float>::PrePack(const Tensor& tensor, int input_idx,
 
 template <typename T>
 Status Gemm<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& /*prepacked_buffers*/,
+                                          gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                           int /*input_idx*/,
                                           /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
@@ -304,6 +305,7 @@ Status Gemm<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& /*prepac
 
 template <>
 Status Gemm<float>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                              gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                               int input_idx,
                                               /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
diff --git a/onnxruntime/core/providers/cpu/math/gemm.h b/onnxruntime/core/providers/cpu/math/gemm.h
index c65f3eb96f62e..d9e66df4bee7c 100644
--- a/onnxruntime/core/providers/cpu/math/gemm.h
+++ b/onnxruntime/core/providers/cpu/math/gemm.h
@@ -37,6 +37,7 @@ class Gemm : protected GemmBase, public OpKernel {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
diff --git a/onnxruntime/core/providers/cpu/math/matmul.cc b/onnxruntime/core/providers/cpu/math/matmul.cc
index 8a7795a81027d..8dea41e3488e2 100644
--- a/onnxruntime/core/providers/cpu/math/matmul.cc
+++ b/onnxruntime/core/providers/cpu/math/matmul.cc
@@ -220,6 +220,7 @@ Status MatMul<float>::PrePack(const Tensor& tensor, int input_idx, /*out*/ Alloc
 }
 
 Status MatMul<float>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                 int input_idx,
                                                 /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
diff --git a/onnxruntime/core/providers/cpu/math/matmul.h b/onnxruntime/core/providers/cpu/math/matmul.h
index 7f2d2ee400b63..9e6ef1a486235 100644
--- a/onnxruntime/core/providers/cpu/math/matmul.h
+++ b/onnxruntime/core/providers/cpu/math/matmul.h
@@ -47,7 +47,9 @@ class MatMul<float> final : public OpKernel {
                  /*out*/ bool& is_packed,
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
-  Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers, int input_idx,
+  Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
+                                   int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
   Status Compute(OpKernelContext* context) const override;
diff --git a/onnxruntime/core/providers/cpu/nn/conv_transpose.cc b/onnxruntime/core/providers/cpu/nn/conv_transpose.cc
index 6ebd12a525371..bbb530d037cec 100644
--- a/onnxruntime/core/providers/cpu/nn/conv_transpose.cc
+++ b/onnxruntime/core/providers/cpu/nn/conv_transpose.cc
@@ -102,6 +102,7 @@ Status ConvTranspose<float>::PrePack(const Tensor& tensor, int input_idx, Alloca
 
 template <typename T>
 Status ConvTranspose<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& /*prepacked_buffers*/,
+                                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                    int /*input_idx*/,
                                                    /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
@@ -110,6 +111,7 @@ Status ConvTranspose<T>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>&
 
 template <>
 Status ConvTranspose<float>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                       gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                        int input_idx,
                                                        /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
diff --git a/onnxruntime/core/providers/cpu/nn/conv_transpose.h b/onnxruntime/core/providers/cpu/nn/conv_transpose.h
index fd6021e65670e..96e3ecf912f32 100644
--- a/onnxruntime/core/providers/cpu/nn/conv_transpose.h
+++ b/onnxruntime/core/providers/cpu/nn/conv_transpose.h
@@ -35,6 +35,7 @@ class ConvTranspose : public OpKernel {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
diff --git a/onnxruntime/core/providers/cpu/quantization/matmul_integer_base.h b/onnxruntime/core/providers/cpu/quantization/matmul_integer_base.h
index fb86e9731035c..9916c426a54fe 100644
--- a/onnxruntime/core/providers/cpu/quantization/matmul_integer_base.h
+++ b/onnxruntime/core/providers/cpu/quantization/matmul_integer_base.h
@@ -80,6 +80,7 @@ class MatMulIntegerBase : public OpKernel {
   }
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override {
     used_shared_buffers = false;
diff --git a/onnxruntime/core/providers/cpu/quantization/qlinearconv.cc b/onnxruntime/core/providers/cpu/quantization/qlinearconv.cc
index 24c8b0d57294e..a5e3d4b04a1e3 100644
--- a/onnxruntime/core/providers/cpu/quantization/qlinearconv.cc
+++ b/onnxruntime/core/providers/cpu/quantization/qlinearconv.cc
@@ -30,6 +30,7 @@ class QLinearConv : public OpKernel {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
@@ -495,6 +496,7 @@ Status QLinearConv<ActType>::PrePack(const Tensor& tensor, int input_idx, Alloca
 
 template <typename ActType>
 Status QLinearConv<ActType>::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                       gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                        int input_idx,
                                                        /*out*/ bool& used_shared_buffers) {
   if (input_idx != 3) {
diff --git a/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.cc b/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.cc
index d1ddd04a953ef..d5be6bd29592e 100644
--- a/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.cc
+++ b/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.cc
@@ -322,6 +322,7 @@ Status DeepCpuGruOp::PrePack(const Tensor& tensor, int input_idx, AllocatorPtr a
 }
 
 Status DeepCpuGruOp::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                               gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                int input_idx,
                                                /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
diff --git a/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.h b/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.h
index 881adf9efb376..fa233cc6f9cde 100644
--- a/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.h
+++ b/onnxruntime/core/providers/cpu/rnn/deep_cpu_gru.h
@@ -69,6 +69,7 @@ class DeepCpuGruOp final : public OpKernel {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
diff --git a/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.cc b/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.cc
index 4b3ea672c0812..d2520804bb64c 100644
--- a/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.cc
+++ b/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.cc
@@ -260,6 +260,7 @@ Status DeepCpuLstmOp::PrePack(const Tensor& tensor, int input_idx,
 }
 
 Status DeepCpuLstmOp::UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                                gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                                 int input_idx,
                                                 /*out*/ bool& used_shared_buffers) {
   used_shared_buffers = false;
diff --git a/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.h b/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.h
index c949b62ce7186..487e2a3fb8129 100644
--- a/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.h
+++ b/onnxruntime/core/providers/cpu/rnn/deep_cpu_lstm.h
@@ -24,6 +24,7 @@ class DeepCpuLstmOp final : public OpKernel, public LSTMBase {
                  /*out*/ PrePackedWeights* prepacked_weights) override;
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override;
 
diff --git a/onnxruntime/core/session/plugin_ep/ep_kernel_registration.cc b/onnxruntime/core/session/plugin_ep/ep_kernel_registration.cc
index 625645e71cfec..6f29361502a73 100644
--- a/onnxruntime/core/session/plugin_ep/ep_kernel_registration.cc
+++ b/onnxruntime/core/session/plugin_ep/ep_kernel_registration.cc
@@ -126,9 +126,9 @@ class PluginEpOpKernel final : public controlflow::IControlFlowKernel {
     return Status::OK();
   }
 
-  Status UseSharedPrePackedBuffers_V2(std::vector<BufferUniquePtr>& buffer_unique_ptrs,
-                                      gsl::span<const size_t> buffer_sizes,
-                                      int input_idx, /*out*/ bool& used_shared_buffers) override {
+  Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& buffer_unique_ptrs,
+                                   gsl::span<const size_t> buffer_sizes,
+                                   int input_idx, /*out*/ bool& used_shared_buffers) override {
     assert(kernel_impl_ != nullptr);  // Should be ensured by PluginEpOpKernel::Create().
 
     if (kernel_impl_->ort_version_supported < 24 || kernel_impl_->SetSharedPrePackedWeight == nullptr) {
diff --git a/onnxruntime/test/framework/session_state_test.cc b/onnxruntime/test/framework/session_state_test.cc
index 656b0ef86289d..418bb2a809259 100644
--- a/onnxruntime/test/framework/session_state_test.cc
+++ b/onnxruntime/test/framework/session_state_test.cc
@@ -662,6 +662,7 @@ class PrePackingTestOpKernel : public OpKernel {
   }
 
   Status UseSharedPrePackedBuffers(std::vector<BufferUniquePtr>& prepacked_buffers,
+                                   gsl::span<const size_t> /*prepacked_buffer_sizes*/,
                                    int input_idx,
                                    /*out*/ bool& used_shared_buffers) override {
     ORT_UNUSED_PARAMETER(input_idx);