diff --git a/ggml-vulkan.cpp b/ggml-vulkan.cpp
index 9c14c0463..9e2846ee4 100644
--- a/ggml-vulkan.cpp
+++ b/ggml-vulkan.cpp
@@ -4534,7 +4534,7 @@ void ggml_vk_graph_cleanup_cpu_assist() {
     ggml_vk_graph_cleanup(ctx);
 }
 
-void ggml_vk_cleanup_cpu_assist() {
+void ggml_vk_free_cpu_assist() {
     ggml_backend_vk_context * ctx = &vk_instance.contexts[0];
 
     if (!ctx->initialized || vk_instance.backends[0] == nullptr) {
diff --git a/ggml-vulkan.h b/ggml-vulkan.h
index 3e1d5dee9..9645126b4 100644
--- a/ggml-vulkan.h
+++ b/ggml-vulkan.h
@@ -20,7 +20,7 @@ GGML_API bool ggml_vk_compute_forward_cpu_assist(struct ggml_compute_params * pa
 void ggml_vk_check_results_1_cpu_assist(struct ggml_compute_params * params, struct ggml_tensor * tensor);
 #endif
 GGML_API void ggml_vk_graph_cleanup_cpu_assist(void);
-GGML_API void ggml_vk_cleanup_cpu_assist(void);
+GGML_API void ggml_vk_free_cpu_assist(void);
 
 // backend API
 GGML_API GGML_CALL ggml_backend_t ggml_backend_vk_init(size_t dev_num);
diff --git a/llama.cpp b/llama.cpp
index e6e9ac1e6..caa4591fc 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -1765,7 +1765,7 @@ struct llama_context {
         }
 
 #ifdef GGML_USE_VULKAN
-        ggml_vk_cleanup_cpu_assist();
+        ggml_vk_free_cpu_assist();
 #endif
 
         ggml_backend_buffer_free(buf_input);