diff --git a/ggml/src/ggml-cuda/vendors/cuda.h b/ggml/src/ggml-cuda/vendors/cuda.h
index 840809a1..30ddbb3c 100644
--- a/ggml/src/ggml-cuda/vendors/cuda.h
+++ b/ggml/src/ggml-cuda/vendors/cuda.h
@@ -6,6 +6,10 @@
 #include <cuda_fp16.h>
 #include <cuda_bf16.h>
 
+#if CUDART_VERSION >= 12050
+#include <cuda_fp8.h>
+#endif // CUDART_VERSION >= 12050
+
 #if CUDART_VERSION < 11020
 #define CU_DEVICE_ATTRIBUTE_VIRTUAL_MEMORY_MANAGEMENT_SUPPORTED CU_DEVICE_ATTRIBUTE_VIRTUAL_ADDRESS_MANAGEMENT_SUPPORTED
 #define CUBLAS_TF32_TENSOR_OP_MATH CUBLAS_TENSOR_OP_MATH